Food Engineering Reviews (2022) 14:134–175

https://doi.org/10.1007/s12393-021-09290-z

Application of Artificial Intelligence in Food Industry—a Guideline

Nidhi Rajesh Mavani1 · Jarinah Mohd Ali1   · Suhaili Othman1,2 · M. A. Hussain3 · Haslaniza Hashim4 ·
Norliza Abd Rahman1

Received: 16 February 2021 / Accepted: 6 July 2021 / Published online: 9 August 2021
© The Author(s) 2021

Abstract
Artificial intelligence (AI) has embodied the recent technology in the food industry over the past few decades due to the
rising of food demands in line with the increasing of the world population. The capability of the said intelligent systems in
various tasks such as food quality determination, control tools, classification of food, and prediction purposes has intensified
their demand in the food industry. Therefore, this paper reviews those diverse applications in comparing their advantages,
limitations, and formulations as a guideline for selecting the most appropriate methods in enhancing future AI- and food
industry–related developments. Furthermore, the integration of this system with other devices such as electronic nose, elec-
tronic tongue, computer vision system, and near infrared spectroscopy (NIR) is also emphasized, all of which will benefit
both the industry players and consumers.

Keywords  AI in food industry · Review · NIR · Food sensors · Model development guidelines

Introduction process industry, food industry, medical industry, data min-

ing, stem cells, and knowledge representation are among
Artificial intelligence (AI) is defined as a field in computer the areas that have been utilizing AI methods [4–11]. AI
science that imitates human thinking processes, learning has a variety of algorithms to choose from such as rein-
ability, and storage of knowledge [1, 2]. AI can be catego- forcement learning, expert system, fuzzy logic (FL), swarm
rized into two types which are strong AI and weak AI. The intelligence, Turing test, cognitive science, artificial neural
weak AI principle is to construct the machine to act as an network (ANN), and logic programming [3]. The alluring
intelligent unit where it mimics the human judgments, while performance of AI has made it the most favorable tool to
the strong AI principle states that the machine can actu- apply in industries including decision making and process
ally represent the human mind [3]. However, strong AI does estimation aiming at overall cost reduction, quality enhance-
not exist yet and the study on this AI is still in progress. ment, and profitability improvement [7, 12].
The gaming industry, weather forecasting, heavy industry, As the population in the world is rising, food demand
is predicted to rise from 59 to 98% by 2050 [13]. Thus, to
cater for this food demand, AI has been applied such as in
* Jarinah Mohd Ali
[email protected]
management of the supply chain, food sorting, production
development, food quality improvement, and proper indus-
1
Department of Chemical and Process Engineering, Faculty trial hygiene [14–16]. Sharma stated that the food processing
of Engineering & Built Environment, Universiti Kebangsaan and handling industries are expected to grow about CAGR
Malaysia, UKM, Selangor 43600 Bangi, Malaysia
of 5% at least until 2021 [15]. ANN has been used as a tool
2
Department of Biological and Agricultural Engineering, in aiding real complex problem solving in the food indus-
Faculty of Engineering, Universiti Putra Malaysia, UPM
Serdang, 43400 Selangor, Malaysia
try according to Funes and coworkers [17], while based on
3
Correa et al., the classification and prediction of parame-
Department of Chemical Engineering, Faculty
of Engineering, University of Malaya, 50603 Kuala Lumpur,
ters are simpler when using ANN, which leads to higher
Malaysia usage demand of ANN over the past years [18]. Besides,
4
Department of Food Sciences, Faculty of Science &
FL and ANN have also acted as controllers in ensuring food
Technology, Universiti Kebangsaan Malaysia, UKM, safety, quality control, yield increment, and production cost
Selangor 43600 Bangi, Malaysia
Food Engineering Reviews (2022) 14:134–175 135

reduction [19, 20]. AI technologies have also known to be in the food industry are illustrated after that section. Finally,
beneficial in food drying technology and as process control a brief conclusion is discussed in this paper.
for the drying process [21–23].
Previous studies have shown many usages of AI in food AI in Food Industry
industries focusing on individual target and aims. A study
has been conducted on the various ANN applications in food The application of AI in the food industry has been growing
process modeling where it has only highlighted the food pro- for years due to various reasons such as food sorting, clas-
cess modeling using ANN [24]. Apart from that, the imple- sification and prediction of the parameters, quality control,
mentation of AI such as ANN, FL, and expert system in food and food safety. Expert system, fuzzy logic, ANN, adaptive
industries have been reviewed but specifically focusing on neuro-fuzzy inference system (ANFIS), and machine learn-
the drying of fresh fruits [23]. A review has been conducted ing are among the popular techniques that have been utilized
on how food safety has been one of the main concerns in the in the food industries. Prior to AI implementation, studies
food industry which leads to the development of smart pack- related to food have been going on over the years to educate
aging systems to fulfill the requirements of the food supply the public about food knowledge as well as to improve the
chain. Intelligent packaging monitors the condition of foods final outcomes related to food properties and the production
to give details on the quality of the food during storage and of foods [31–36]. A lot of benefits can be obtained by using
transportation [25]. Another study reviewed on intelligent the AI method, and its implementation in the food industry
packaging as a tool to minimize food waste where about 45 has been going on since decades ago and has been increas-
recent advances in the field of optical systems for freshness ing till today [37–39, 31, 32]. Nevertheless, this paper will
monitoring have been reported. Meat, fish products, fruits, focus on the application of AI in food industries from the
and vegetables were covered in the study as they are the year 2015 onwards since tremendous increase and innova-
most representative fields of application [25]. Few different tion are seen in the implementation recently. It is worth not-
studies have been conducted on intelligent packaging, and ing that several methods such as partial least square, gastro-
these studies proved that the usage of intelligent packaging intestinal unified theoretical framework, in silico models,
systems plays an important role in the food factory in the empirical models, sparse regression, successive projections
context of the food chain as they are able to monitor the algorithms, and competitive adaptive reweighted sampling
freshness of food products and crops [23, 26–30]. which have been used for prediction and enhancement of the
There are also several other studies that have been con- food industries are not discussed here; instead it is narrowed
ducted on the application of AI and sensors in food; how- down to the wide application of AI in the food industry.
ever, the coverage is rather limited. Therefore, a comprehen-
sive review that assembles all AI applications in the food Knowledge‑based Expert System in Food Industry
industry as well as its combinations with appropriate sensor
will be a great advantage, all of which are unavailable as The knowledge-based system is a computer program that
to the knowledge of the author. Such review will assist in utilizes knowledge from different sources, information, and
gathering the advantages, limitations, and methodologies as data to solve complicated problems. It can be classified into
a one-stop guideline and reference for food industry play- three categories which are expert systems, knowledge-based
ers, practitioners, and academicians. To be exact, different artificial intelligence, and knowledge-based engineering. The
types of AI and their recent application in food industries breakdown of the knowledge-based system is presented in
will be highlighted which comprises several AI techniques Fig. 1. The knowledge-based expert system which is widely
including expert system, fuzzy logic, ANN, and machine used in the industries is a decisive and collective computer
learning. In addition, the integration of AI with electronic system that is able to imitate the decision-making ability
nose (E-nose), electronic tongue (E-tongue), near infrared of human expert [40]. It is a type of knowledge-based sys-
spectroscopy (NIRS), and computer vision system (CVS) tem that is known as among the first successful AI models.
is also provided. This paper is organized as follows. The This system depends on experts for solving the complicated
introduction of AI is explained in the first section followed issues in a particular domain. It has two sub-systems, which
by the application of different types of AI in the food indus- are knowledge base and inference engine. The facts about
try. Following that, the fusion of the AI with the external the world are stored in the knowledge base, and the infer-
sensors in the food industry is presented. In the latter part, ence engine represents the rules and conditions regarding the
a critical review is conducted where discussion on the main world which are usually expressed in terms of the IF–THEN
application of the AI algorithms in the food industry is car- rules [41]. Normally, it is able to resolve complicated issues
ried out. A flowchart is presented to assist the researchers on by the aid of a human expert. This system is based on the
establishing the most appropriate AI model based on their knowledge from the experts. The main components of the
specific case study. Then, the trends on the application of AI expert system (ES) are human expert, knowledge engineer,
136 Food Engineering Reviews (2022) 14:134–175

Fig. 1  Knowledge-based system
Knowledge-based Systems

Knowledge Base AI Expert System Knowledge-based Engineering

Statistical Corpora Knowledge Bases

knowledge base, inference engine, user interface, and the Fuzzy Logic Technique in the Food Industry
user. The flow of the expert system is shown in Fig. 2.
The food industry has been utilizing ES for various objec- Fuzzy logic (FL) was first introduced by Zadeh in 1965
tives as this system is proven to be useful especially in the based on the impeccable capability of human intellect in
decision-making process. The knowledge-based expert decision making and unraveling the imprecise, uncertain,
system has been applied in white winemaking during the and ambiguous data while solving problems [47, 48]. The
fermentation process for the supervision, intelligent con- fuzzy set theory is recognized in such a manner that an ele-
trol, and data recovery [42]. Apart from that, a web-based ment belongs to a fuzzy set with a certain degree of mem-
application was developed by implementing the ES to cal- bership which has a real number in the interval [0, 1] [49].
culate the nutritional value of the food for the users, and the FL models consist of several steps which are fuzzification,
development of ES was able to help the SMIs in obtaining inference system, and defuzzification process [50,  51].
the details required for the qualification in obtaining the food Fuzzification is a process where the crisp value is converted
production certificates [43]. Food safety is very important into a degree of membership and yields the fuzzy input sets.
in the food industry,thus, the application of ES that is linked The corresponding degree in the membership functions is
closely to food safety has been used extensively ranging normally between 0 and 1. [52]. There are a variety of mem-
from process design, safety management, quality of food, bership functions to choose from, whereby the commonly
and risk assessment [44]. Furthermore, a prototype informa- used ones are triangular, Z-shaped, S-shaped, trapezoidal,
tion technology tool and guidelines with corrective actions and Gaussian-shaped [52]. The inference system is where
that considered ES in the model were developed for the food the fuzzy input is being translated to get output by using
industry where few essential factors such as food safety, the fuzzy rules. The fuzzy rules are known as IF–THEN
nutrition, quality, and cost were studied [45]. In addition, a rules where it is written such IF premise, THEN consequent
digital learning tool, namely, MESTRAL, was developed to whereby the IF comprises input parameters and THEN is the
assist people in food processing by using models developed output parameters [53]. The inference system consists of the
from research in food science and technology and simula- style which is either the Mamdani or Takagi–Sugeno Kang
tors. This tool is based on the knowledge engineering and (TSK). Defuzzification is the ultimate phase in the fuzzy
reflected real applications which can be mapped with the logic model where the crisp values are obtained [54]. There
system scale and knowledge frameworks [46]. A compre- are different methods of defuzzification which are center of
hensive review was conducted by Leo Kumar on the applica- gravity, mean of maximum, smallest of maximum, largest
tion of the knowledge-based expert system in manufacturing of maximum, center of maximum, and centroid of area [55].
planning. The paper has also discussed the utilization of ES FL has been long utilized in the industry due to its sim-
in decision making in three wide areas which are the process plicity and ability to solve problems in a fast and accurate
planning activities, diverse applications, and manufacturing manner. FL has been employed in the food industry in food
planning [41]. Moreover, Table 1 gathers some of the recent modeling, control, and classification and in addressing food-
application of ES in the food industry ranging from the raw related problems by managing human reasoning in linguistic
material to the final production as well as the food safety. terms [56]. The food manufacturing system has improved

Fig. 2  Expert system
Human Expert Knowledge Engineer Knowledge Base

Inference Engine

User Interface User

Table 1  Application of expert system in food industry
Application Objective Category in the industry Important outcomes References

Banana To identify the banana disease and methods to Raw material/agriculture (i) The system was able to diagnose the diseases Budiyanto et al. [133]
overcome them in the plants based on the stems, leaves, and
roots which assists in preventing from the
diseases happening
Barley To classify the barley grains Quality control (i) The classification accuracy by using the ES Szturo & Szczypinski [40]
method was above 72% which was considered
satisfactory
Coffee To determine the appropriate process on the dry Production/quality control (i) ES was able to select the correct production Hernández-Vera et al. [134]
mill process for a specific coffee with an accuracy
of 93%
Food Engineering Reviews (2022) 14:134–175

Coffee beans To identify the quality grading of the coffee beans Sensory evaluation (i) The developed fuzzy ES, namely, AI Cupper, Livio & Hodhod [135]
was able to grade the coffee with an accuracy
of 95%
Corn To detect the corn pests and diseases Agriculture/raw material (i) The system was able to detect the pests and Sumaryanti et al. [136]
disease with an accuracy of 76.6% and also
provide ways to control them
(ii) The developed system was able to provide
facility explanation regarding the diagnosis
results
Food additives To identify the halal safety rating for food addi- Food dafety (i) The developed Halal Food Additive using ES Zakaria et al. [137]
tives provides the consumers with safety rating key
according to their past food consumption record
experience
(ii) It provides consumers in finding nutritional
foods and at the same time fulfilling the
Halalan-Toyyiban criteria
Food products To monitor and forecast the product quality in the Production/food quality (i) An intelligent ES was developed which is able Blagoveshchenskiy et al.
production process to monitor the product quality indicators and [138]
make changes to the existing methods; recom-
mendations for the products and the defects for
the final product can be identified by this system
Fresh food To optimize the distribution networks of the fresh Sustainability (i) The proposed expert system, Food Distribu- Bortolini et al. [139]
foods tion Planner, was able to develop the most
effective distribution method which reduces the
emission of carbon dioxide by 9.6%, increase of
2.7% in operating cost, and no waste produced
during the delivery time due to the preservation
method utilized during the shipment
137
Table 1  (continued)
138

Application Objective Category in the industry Important outcomes References

Livestock production To analyze the outcome of various variables on Production/raw material (i) ES can be utilized as a decision support system Vásquez et al. [140]
(milk, meat) the performance of the livestock production for livestock producers for identifying the best
practices for the livestock which maximizes the
production of the meat and milk
(ii) The greatest impact on the production is the
type of grazing being fed to the cattle as the diet
affects the health of the cow who produces the
milk and the meat
Red wine and rum To forecast the key aroma compounds for foods Sensory evaluation/quality control (i) The developed rapid method system, Sensom- Nicolotti et al. [141]
without using the human olfactory system ics Based Expert System, resulted in a good
agreement in the key odorants for the food
aroma distillate
Rice crops To aid the farmers in making decisions for the Agriculture/ (i) The farmers are assisted by the ES in selecting Kharisma et al. [142]
rice crops production the seeds and tackling the pests and diseases for
the rice crops which eventually will improve the
production of the rice crops
Soybean To identify the diseases on soybeans Agriculture (i) The study successfully designed an ES to Rajendra et al. [143]
identify soybean diseases by comparing the
accuracy using the frame-based representation
and rule-based representation method
(ii) Frame-based representation ES has shown a
higher accuracy compared to the rule-based ES
White winemaking To develop a knowledge-based ES for the Processing/sustainability (i) A cost-efficiency advanced control system Sipos [42]
alcoholic fermentation process of the white through the knowledge-based ES was developed
winemaking for the alcoholic fermentation process which
was used for the supervision, control, and data
recovery software of the bioreactor
(ii) It was proven to be applied in winemaking at
the industrial scale and can be adjusted for few
areas in the food manufacturing sectors
Wine To measure the environmental impact of viticul- Sustainability (i) ES integrated with the geographic information Lamastra et al. [144]
ture at wine estate system software was able to measure the envi-
ronmental impact of viticulture in a comprehen-
sive way
(ii) The model is said to be an environmental sup-
port system in supporting policy and decision-
making in the management
Food Engineering Reviews (2022) 14:134–175
Food Engineering Reviews (2022) 14:134–175 139

by the implementation of the fuzzy logic where about 7% output data can be calculated by using the equation shown
of electricity losses has been reduced compared to the con- based on Fig. 4.
ventional regulation method [57]. Sensory evaluation of the Previous studies have highlighted the utilization of ANN
food is also one of the most common parts where FL plays in numerous tasks within the food industry. This includes the
an important role. Furthermore, a quicker solution to prob- assessment and classification of the samples, complex cal-
lems can be performed by using a system involving fuzzy culation such as heat and mass transfer, and analysis of the
rules [58]. Table 2 shows previous applications of FL in the existing data for control purposes as well as for prediction
food industry and their attributes. From a previous study, FL purposes which are listed in Table 3. All applications have
has been proven to successfully maintain the quality of the shown satisfactory performances based on the R2 values,
foods, and it acts as a prediction tool and control system for showing that ANN can provide results in an accurate and
food production processes. reliable manner.

ANN Technique in  the  Food Industry  ANN is another AI Machine Learning Techniques
element, which is also commonly applied in the food indus-
try. ANN is designed to mimic the human brain and be able Machine learning (ML) is known to be the subset of AI
to gain knowledge through learning and the inter-neuro con- [73, 74]. It is a computer algorithm that advances auto-
nections which are known as synaptic weights [59, 60]. Gan- matically with experiences. ML can be classified into three
dhi and coworkers have stated that the configuration of ANN broad categories which are supervised learning, unsuper-
is designed in such a way that it will accommodate certain vised learning, and reinforcement learning [11, 75]. Super-
application such as data classification or pattern recognition vised learning aims to predict the desired target or output
[61]. According to Gonzalez-Fernandez, ANN is applicable by applying the given set of inputs [76]. On the other hand,
to a different kind of problems and situations, adaptable, and unsupervised learning does not have any outputs to be pre-
flexible. In addition, Gonzalez et al. (2019) have also stated dicted and this method is utilized to classify the given data
that ANN is suitable to model most non-linear systems and and determine the naturally occurring patterns [77]. Rein-
is adaptable to new situations even though adjustments are forcement learning is when there is an interaction between
needed. Moreover, the most outstanding features of ANN the program and the environment in reaching certain goals
is its non-linear regression [62]. There are several types of [78]. Among the known models in machine learning are
ANN including feedforward neural network, radial basis ANN, decision trees (DT), support vector machines (SVM),
function neural network, Kohonen self-organizing neural regression analysis, Bayesian networks, genetic algorithm,
network, recurrent neural network, convolutional neural net- kernel machines, and federated learning [76, 79]. ML has
work, and modular neural network [63]. Multilayer percep- been commonly used for handling complex tasks and huge
tron (MLP), radial basis function networks (RBFNN), and amount of data as well as variety of variables where no pre-
Kohonen self-organizing algorithms are the most effective formula or existing formula is available for the problem.
types of NN when it comes to solving real problems [61]. Other than that, ML models have the additional ability to
The most common network that is used for prediction and learn from examples instead of being programmed with rules
pattern recognition is the multilayer perceptron [18, 64, 65]. [80].
Besides that, ANN learning could be classified into super- Among the ML methods that are used in the food industry
vised and unsupervised depending on the learning tech- include ordinary least square regression (OLS-R), stepwise
niques [17]. In general, the structure of ANN consisted of linear regression (SL-R), principal component regression
an input layer, hidden layer, and output layer, either single or (PC-R), partial least square regression (PLS-R), support vec-
many layers [66–68]. The architecture comprises activation tor regression (SVM-R), boosted logistic regression (BLR),
functions, namely, the feed-forward or feedback [69]. The random forest regression (RF-R), and k-nearest neighbors’
backpropagation learning algorithm is normally used as it regression (kNN-R) [81]. Studies showed that the usage of
is able to minimize the prediction error by feeding it back ML has helped in reducing the sensory evaluation cost, in
as an input until the minimum acceptable error is obtained decision making, and in enhancing business strategies so as
[18]. An additional input known as bias is added to neurons to cater users’ need [82]. Long short-term memory (LSTM)
which allows a portrayal of phenomena having thresholds which is an artificial recurrent neural network has been
[70, 71]. In ANN, the dataset is normally associated with a employed in the food industry as pH detection in the cheese
learning algorithm which trained the network and could be fermentation process [83]. On the other hand, GA has been
categorized into three groups specifically supervised, unsu- utilized for finding the optimum parameters in food whereas
pervised, and reinforcement learning [72]. Then, the data NN has been occupied to predict the final fouling rate in
will undergo training and testing for analyzing the outputs. food processing [84]. ML has shown to be advantageous in
The general structure for the ANN is shown in Fig. 3. The predicting the food insecurity in the UK [85]. Apart from
Table 2  Application of fuzzy logic in the food industry
140

Application Objective Fuzzy Membership function Important outcomes References

inference
system

Aromatic foods To rank the sensory attributes of aromatic foods Mamdani Triangular (i) Overall ranking for tea and tastemaker and Chowdhury & Das [145]
packed in films made from corn starch the important quality attributes of the food
materials in general, and the samples were able
to be done using FL
(ii) Aroma and taste of tea leaf and tastemaker in
general were assessed as “Highly important”
sensory attributes
Beetroot candy To rank the candy with various content ratios Mamdani Triangular (i) The developed model was able to optimize Fatma et al. [146]
and perform the ranking of the candy involv-
ing different ingredient ratios
Canned food To control sterilization temperature by using Mamdani Triangular (i) The sterilizing temperature with an accu- Chung et al. [147]
fuzzy logic and making online corrections in racy of ± 0.5 ℃ can be maintained by a fuzzy
autoclave operation controller
(ii) Batch processing can be completed using
the proposed system with less time, steam
consumption, and risk of over-sterilization
Coffee To determine the suitable process on a dry mill Mamdani Triangular (i) The developed system will be useful for the Hernández-Vera et al. [148]
according to customer requirements using an correct decision process between two different
expert system based on fuzzy logic types of coffee
(ii) Validation was carried out by comparing
the process values by the model with the real
process data, and the coefficient of determina-
tion obtained was 93%
Coffee beans To introduce a control system for the roasting Mamdani Triangular (i) The consistent roasting level of the beans is Harsawardana et al. [149]
machine able to be produced by the proposed model
Cupcakes To rank the cupcakes according to their quality Mamdani Triangular (i) The system was able to determine the best Singh et al. [150]
attributes condition for their cupcakes with respect to
their sensory attributes
(ii) The ranking of the quality attributes was able
to be performed by the system
Dough To implement the FL to act as a controller Mamdani Triangular, trapezoi- (i) The settling time and the response of the FL Yousefi-Darani et al. [151]
system in bread making dal controller showed a better performance than
the proportional-integral-derivative (PID)
controller
(ii) The FL controller system was established
successfully for the proofing process in bread
making
Food Engineering Reviews (2022) 14:134–175
Table 2  (continued)
Application Objective Fuzzy Membership function Important outcomes References
inference
system

Fava beans To predict the physical parameters of the beans Mamdani Triangular (i) The model was able to predict thirteen Farzaneh et al. (152, 153
with various moisture contents parameters of the beans with the moisture con-
tent ranging from 9.3 to 31.3% in the input
(ii) Comparison was done between the FL
results and the experimental value where a
high correlation value which is 0.999 and
mean standard deviation ranging 1.23–12.56%
were obtained as an effective system and can
be used to develop a model for controlling and
Food Engineering Reviews (2022) 14:134–175

managing various stages during processing

Flixweed seed To rank different ways of extraction in prepar- Mamdani Triangular (i) Ranking based on the properties, extraction Shahidi et al. [154]
ing the seeds output, and duration was able to determine the
best method for the preparation of the seeds
Fresh mango juice To study the effect of high pressure processes Mamdani Triangular (i) The model was able to determine that the Kaushik et al. [155]
(litchi juice) (HPP) on sensory attributes of fresh mango HPP effect depended on the type of products
juice and litchi juice and domain of pressure–temperature
(ii) HPP is proven to be a promising method for
the preservation of fruit products
Hydrogel colloidosomes To estimate the release of caffeine from hydro- Mamdani Type S, type Z (i) The proposed diffusional-fuzzy model able Amiryousefi et al. [156]
gel colloidosome to describe the caffeine release from hydrogel
colloidosomes
(ii) The model has a higher precision, better
handling of uncertainty property and better
generalization capability
Onions To predict the drying kinetics of the onions Mamdani Triangular (i) The model was able to predict the moisture Jafari et al. [157]
ratio at varying conditions with high perfor-
mance where the value of R obtained was
0.9999 and the low root mean square error
(RMSE) was 0.004157
Pineapple Rasgulla To rank the pineapple Rasgulla with respect to TSK Triangular (i) Sensory evaluation for different concentra- Sarkar et al. [158]
the parameters tions of pineapple Rasgulla and ranking of the
samples was performed successfully by using
the FL model
Pizza production industry To develop a system in order to improve the Mamdani Triangular (i) The developed FL control system is able to Blasi [159]
production system identify the amounts of workers and ovens
needed in pizza production which improves the
customer’s satisfactory level by reducing the
waiting time as well as reducing the wastage
Salt To estimate the production of salt by variables TSK Triangular (i) By using the Sugeno zero-order model, the Yulianto et al. [160]
that affect it time for production of salt could be estimated
with a minor error value of 0.0917
141
 142 Food Engineering Reviews (2022) 14:134–175

that, ML has also proven to have predicted the trend of sales

in the food industry [86] In addition to that, ML was also

Mahadevappa et al. [12]

Jahedi Rad et al. [163]

Zare & Ghazali [161]
able to predict the food waste generated and give an insight
to the production system [87]. Major applications of ML
in the food industry and its positive highlights are briefly

Basak [162]
References

emphasized in Table 4.

Adaptive Neuro Fuzzy Inference System (ANFIS) Techniques

the dough which improves the sheeting process

(i) Able to decide the best program with the least
(ii) The Pearson correlation r value obtained was

number of rolling steps based on the quality of

customers’ acceptability on treated apple juice
(i) The model was able to determine the quality

high accuracy where the value of R2 equals to

ANFIS is a type of AI where FL and ANN are combined

of mulberry under varying conditions with a

(i) The model could predict the moisture ratio
analysis gave an insight into variation into
greater than 0.95 at different temperatures

in such a way that it integrates the human-like reasoning

(i) The FL approach by applying similarity
of the fish at the initial stages of storage

style of the FL system with the computational and learning

0.9996 and RMSE value of 0.01095

capabilities of ANN [56]. In ANFIS, the learning procedure
is transferred from the neural network into the FL system
where a set of fuzzy rules with suitable membership func-
tions from the data obtained is developed [88]. Mamat et al.
[89] stated that uncertainty data could be processed and
Membership function Important outcomes

gain higher accuracy when ANFIS is applied [89]. Besides,

of the dough

ANFIS is also known as a fast and robust method in solving

problems [90]. Not only that, Sharma et al. [91] also claimed
that ANFIS has a higher performance compared to other
models such as ANN and multiple regression models in their
study [91]. ANFIS is a fuzzy reasoning system and combi-
nation of the parameters trained by ANN-based algorithms.
The fuzzy inference system that is normally used is Takagi
Trapezoidal

Sugeno Kang in the ANFIS model with the feedforward

Triangular

Triangular

Triangular

neural network consisting of the learning algorithms [92].

The structure of ANFIS is made up of five layers which are
fuzzy layer, product layer, normalized layer, defuzzification
layer, and total output layer [93, 31, 32]. The backpropaga-
Mamdani

Mamdani

Mamdani

Mamdani
inference
system

tion algorithm has been normally applied in the model in

Fuzzy

order to avoid over-fitting from occurring [92]. A high cor-

relation value (R2) indicates that the developed model has
To assess fish quality by biogenic amines using

To carry out a sensory evaluation of raw apple

juice treated with raw betel leaf essential oil

To forecast the moisture ratio of the mulberry

high accuracy and is suitable for industrial applications. The

general structure of the ANFIS model is illustrated in Fig. 5.
To develop an effective sheeting of wheat

The first layer in ANFIS has nodes that are adjustable,

and it is called as the premise parameters [56]. The second
layer in ANFIS has fixed nodes, and the output is the product
during the drying process

of all incoming signals. Every output node represents the

firing strength of the rule. The third layer consists of fixed
a fuzzy logic model

node labeled as N. The outputs of the third layer are called

normalized firing strengths. Every node in the fourth layer
is an adaptive node with a node function, and the parameters
Objective

in this layer are called as the subsequent parameters [56].

dough

The final layer in the ANFIS layer has a fixed single node
which calculates the overall output as the summation of all
the incoming signals. The calculation involved in each layer
Treated raw apple juice

is shown below. The output of the ith model in layer 1 is

Table 2  (continued)

denoted as ­01, i.
White mulberry

Layer 1: O1,i = 𝜇Ai (x), fori = 1,2 atau O1,i = 𝜇Bi−2 (y),
Wheat dough
Application

fori = 3,4.
Sardine

Layer 2: O2,i = wi = 𝜇Ai (x)𝜇Bi (y), fori = 1,2.

Layer 3: O3,i = w = w +w , i = 1,2.
wi
1 2
Food Engineering Reviews (2022) 14:134–175 143

Fig. 3  ANN structure in general

Input Output

Output Layer
Hidden Layer

Input Layer

Layer 4: O4,i = wfi = wi (pi x + qi y + ri ); wi is the normal- Integrating AI with External Sensors for Real‑time
ized firing strength from layer 3 and. Detection in Food Industry
{ pi , qi , ri } is the parameter∑set of this node.
FL or ANN is often integrated with several sensors for real-
wf
Layer 5: O5,1 = i wi fi = ∑i wi i .
∑
time detection such as electronic nose (E-nose), electronic
i i
The ANFIS model is attractive enough that it could
solve problems related to the food industry, which are tongue (E-tongue), machine learning (ML), computer vision
complicated, practical, and barely solved by other meth- system (CVS), and near infrared spectroscopy (NIRS) for
ods and has been widely used in the food industry for real-time detection and to obtain higher accuracy results
prediction and classification purposes. ANFIS has been in a shorter time. These detectors have also combined their
applied in various food processing involving recent tech- elements together for enhancing their accuracy and targeted
nology which comprised five main categories which are results. The integration of these sensors with the artificial
food property prediction, drying of food, thermal process intelligence methods has been shown quite a number in food
modeling, microbial growth, and quality control of food industries over the past few years.
as well as food rheology [56]. The utilization of ANFIS Electronic nose also known as E-nose is an instrument cre-
in the food industry has been commenced years ago, and ated to sense odors or flavors in analogy to the human nose.
Table 5 describes those applications. It consists of an array of electronic chemical sensors where
it is able to recognize both simple and complex odors [94].
E-nose has been used in gas sensing where the analysis of each
component or mixture of gases/vapors is required. Besides, it
plays an important role in the food industry for controlling the
quality of the products. Due to its ability to detect complex

Fig. 4  General calculation in w1
ANN
X1 Activation function
w2

X2 ∑⬚ F Y
Y in

Xm w3

Total input, ( ) = ∑ .
Total output, Y = F(Yin)
Table 3  Application of ANN in the food industry
144

Application Objectives Types of ANN Outcomes References

Cocoa powder To predict the effect of process parameters on the proper- MLP (i) The model could predict the effect of different param- Benković et al. [164]
ties of cocoa mixtures eters changes on the physical and chemical properties of
the cocoa mixtures with a high accuracy where the value
of R2 derived was 0.934 with low error of 0.053
Dried vegetables To identify the quality of dried carrots and classification of MLP (i) Suitable to be used for assessment and classification of Koszela et al. [165]
such dried products dry carrots
(ii) Enable the selection of important representative char-
acteristics for the quality assessment
(iii) Comparison between algorithms was done, and the
backpropagation is the most successful algorithm com-
pared to other algorithms
Eggplant To describe the mass transfer kinetics of eggplants in MLP (i) High volume of complicated problems is able to be Bahmani et al. [166]
osmotic dehydration (OD) modeled and analyzed by using ANN, and it is the most
suitable software for calculation problems
(ii) The developed model achieved the highest R2 value of
0.9825
Extra virgin olive oils To evaluate the influence of light exposure conditions and MLP (i) ANN showed a high classification performance with an S. F. Silva et al. (71, 70
packaging material on the stability of physicochemical accuracy of greater than 90% for the test data and greater
characteristics of extra virgin olive oils than 85% for the training set
(ii) Showed the robustness of the model which indicates
the suitability for solving clustering, pattern recognition,
classification, and adulteration issues regarding extra
virgin olive oils
Garlic To forecast the sensory quality of garlic MLP (i) A model with the best prediction of the sensory quality Liu et al. [167]
of the garlic and garlic products was developed with an
R2 value of 0.9866
Honey To predict the stability of the Indian honey crystallization MLP (i) The ANN model was able to forecast the stability of Naik et al. [168]
of different component ratios the honey consisting of different compositions with high
accuracy where the R value obtained was 0.9994
Mango To estimate the weight of the mango MLP (i) Estimation of the weight of the mangoes was able to Dang et al. [169]
be done by applying thirteen different parameters in the
model
Mushrooms To develop an ANN model that can predict the moisture MLP (i) The ANN model was able to predict the moisture con- Omari et al. [170]
content of the mushrooms during the drying process tent of the mushrooms during the drying process with an
R value of 0.9817
Mushrooms To predict the temperature variety of mushroom grow- MLP, RBF (i) The prediction by using the RBF method has a higher Ardabili et al. [171]
ing hall based on the parameters affecting the room accuracy compared to that of MLP where the value of
temperature correlation achieved was 0.996 and 0.9612, respectively
Onions To estimate the drying behavior of onion MLP (i) The develop model was able to forecast the drying Jafari et al. [157]
kinetics of onions at different temperatures and times
with a high performance where the R value achieved a
value of 0.99956
Food Engineering Reviews (2022) 14:134–175
Table 3  (continued)
Application Objectives Types of ANN Outcomes References

Potato cubes To carry out analysis in a fluidized bed dryer for the drying MLP (i) The developed model was able to carry out the analysis Azadbakht et al. [172]
of potato cubes under different conditions of energy and exergy losses in the dryer for the drying of
the potato cubes
(ii) All the analysis done by using the ANN model
obtained a high value of R2 which is greater than 0.98
and the average value obtained was 0.99
Potato peels To model a system that can predict and optimize the MLP (i) This study with the aid of the ANN model actually Anastácio et al. [173]
extracted conditions by using the response surface meth- helped to determine that the potato peels which are often
odology and ANN thrown away as a waste are actually beneficial
(ii) The R2 value achieved by the ANN model for the
analysis of three different extracted values were greater
Food Engineering Reviews (2022) 14:134–175

than 0.93
Quince fruit To determine the moisture ratio of the fruit during the MLP (i) By implementing different ANN structures, the Chasiotis et al. [174]
drying process and test the performance of the developed moisture content ratio during the drying process for the
model quince fruits was modeled successfully
(ii) The developed model was able to predict the moisture
ratio with high correlation value where the R2 obtained
was greater than 99%
Rice crop To predict rice production yield and investigate the factors MLP (i) A good alternative to predict the rice production yield Gandhi et al. [61]
affecting the rice crop yield compared to traditional linear regression methods
(ii) The accuracy obtained by the developed algorithm was
97.54%, with the sensitivity of 96.33% and specificity of
98.12%
Sausage To forecast the benzo[a]pyrene (BaP) content of smoked MLP (i) The model was able to predict the BaP content in Chen et al. [175]
sausage smoked sausages and create a control system for the
smoking to reduce the BaP contamination in smoked
sausages
(ii) The model has a high accuracy where the overall pre-
diction was greater than 0.90
Vegetable oils To classify vegetable oils: canola, sunflower, corn, and MLP (i) Able to simplify the vegetable oil classification with Silva et al. [71], Da et al. [70]
soybean using a very few mathematical manipulation and high accuracy
ANN (ii) A fast-network training and uses very few mathemati-
cal manipulations in the spectra data
145
146 Food Engineering Reviews (2022) 14:134–175

Table 4  Application of machine learning in food industries

Application ML methods Important outcomes References

Apple Linear discriminant analysis, adaptive

(i) The ML was able to classify the Li et al. [176]
boosting apples accurately with a rate of
100% using the collected acoustic
emission signals
Artichoke MLP, RF, BLR (i) The characteristics of ion patterns Sabater et al. [177]
were done by the ML models to be
set up for each enzyme with a high
prediction rate of 95% above
Beer ANN (i) The ANN model was able to clas- Claudia Gonzalez et al. [178]
sify the chemical components in the
beer with a high overall accuracy
of 95%
Biscuits Convolutional neural network (i) The developed model was able to De Sousa Silva et al. [179]
classify and evaluate the quality of
different types of biscuits with an
accuracy up to 99%
Cheese LSTM (i) The combination of mechanistic Li et al. [83]
modeling with LSTM was able to
describe the changes in lactic acid,
lactose, and biomass with high accu-
racy where the value of R2 obtained
is greater than 0.99
(ii) The prediction of pH for the
cheese fermentation is able to be
done by using the developed model
Citrus limetta (Mosambi peel) SVM-ANN, SVM-Gaussian process (i) The SVM model was able to Younis et al. [180]
regression (SVM-GPR) predict and classify all the results
for the taste of lime powder that
has been treated by the salt with an
accuracy of 1.0
(ii) Optimization by using the ML
tool allows to maintain the taste and
retain the polyphenol content in the
lime
Fruits (Arbutus unedo L. fruits) RF, SVM, ANN (i) The stability of the extracts in the Astray et al. [181]
Arbutus unedo L. fruits in terms of
aqueous and powder systems was
able to be done by the ML methods
with overall coefficient in the range
of 0.9128 and 0.9912 for the best
models chosen
Lamb meat SVM (i) The classification accuracy of the Alaiz-Rodriguez & Parnell
lamb meat fat increased from 89.70 [85]
to 93.89% by using the SVM method
Mangoes Naive Bayes, SVM (i) The system is able to detect the Pise & Upadhye [182]
maturity of the mangoes based on
their quality attributes
Meat OLS-R, SL-R, PC-R, PLS-R, SVM-R, (i) Different kinds of microorganisms Estelles-Lopez et al. [81]
RF-R, and kNN-R causing the beef spoilage could be
detected by using the regression ML
that obtained the data from five dif-
ferent analytical methods
(ii) All the methods were able to pre-
dict in all cases accurately with the
rank order of RF-R, PLS-R, kNN-R,
PC-R, and SVM-R
Food Engineering Reviews (2022) 14:134–175 147

Table 4  (continued)
Application ML methods Important outcomes References

Milk SVM (i) Presence and the level of antibiot- Gutiérrez et al. [183]
ics concentration in the cow milk
was determined by using the SVM
classifiers with high accuracy rate
of greater than 83% and greater
sensitivity compared to the typical
metrics
Salmon TreeBagger (i) The established model was able to Xu & Sun [184]
classify the normal and freezer burnt
categories with high accuracy where
the correct classification rate yielded
0.914 for validation and 0.978 in
cross validation
Wine SVM, RF, MLP (i) The comparison among the Shaw et al. [185]
three algorithms were done in the
evaluation of wine quality and the
best result was obtained by the RF
method with an average accuracy of
81.96% meanwhile others delivered
a low accuracy late. This indicates
the RF algorithm can be used to
evaluate the quality of the wine

odors, it has been employed as an environment protection tool the methodology employed in E-nose [96]. The general system
and detection of explosives materials [95]. An array of non- of E-nose is shown in Fig. 6.
specific gas sensors is known to be the main hardware com- E-nose has been widely used to aid in both quality control
ponent of E-nose where the sensors will interact with a variety and assurance in the food industries. Wines, grains, cooking oils,
of chemicals with differing strengths. It then stimulates the eggs, dairy products, meat and dairy products, meat, fish prod-
sensors in the array where characteristic response is extracted ucts, fresh-cut and processed vegetables, tea, coffee, and juices
known as a fingerprint [94]. The main software component of have successfully applied e-nose for sampling classification,
E-nose is its feature extraction and pattern recognition algo- detection, and quality control. E-nose has successfully classi-
rithms where the response is processed, important details are fied samples with different molecular compounds [97]. Besides,
elicited and then chosen. Thus, the software component of Sanaeifar et al. have reviewed and confirmed that e-nose was
the E-nose is greatly important to stimulate its performance. able to detect defects and contamination in foodstuffs [98].
In general, E-nose is divided into three main parts, namely, Classification and differentiation of different fruits have also
sample delivery system, a detection system, and a computing determined by using e-nose [99]. A review has been conducted
system. ANN, FL, and pattern recognitions are the examples of on the application of the E-nose for monitoring the authenticity

Fig. 5  General structure of x y
ANFIS

w1
A1
x
A2 f

B1
y
w2
B2 x y

Layer 1 Layer 2 Layer 3 Layer 4 Layer 5

148 Food Engineering Reviews (2022) 14:134–175

of food [100]. Adding to this, Mohamed et al. have carried out Electronic tongue (E-tongue) is an instrument that is able
a comprehensive review on the classification of food freshness to determine and analyze taste. Several low-selective sensors
by using e-nose integrated with the FL and ANN method [101]. are available in E-tongue which is also known as “a multi-
Recent application of e-nose with computing methods involving sensory system,” and advanced mathematical technique is
AI in food industries is shown in Table 6. being used to process the signal based on pattern recognition

Table 5  Application of ANFIS in the food industry

Applications Outcomes References

Fish oil (i) A model was developed to estimate the oxidation parameters using three Asnaashari et al. [186]
different algorithms which are, ANFIS, multilinear perceptron, and multiple
linear regression, and it was found that ANFIS model had the best accuracy in
predicting the parameters
Ice cream (i) The sensory attributes of ice cream were investigated by using the ANFIS Bahram-Parvar et al. [187]
model to predict the acceptability of taste with respect to the input parameters
and the model achieved a minimum error of 5.11% and high correlation value
of 0.93
Indian sweets (Pantoa) (i) The prediction of the heat transfer coefficient during the frying of pantoa Neethu et al. [188]
using the ANFIS model yielded a high R2 value of 0.9984, and this prediction
is important for designing the process equipment as well as saving energy in
commercial production
Orange (i) The developed ANFIS model was able to predict the orange taste and has Mokarram et al. [189]
higher performance when compared to multiple regression model
Quince fruits (i) The ANFIS model was able to predict the moisture ratio, energy utilization, Abbaspour-Gilandeh et al. [190]
energy utilization ratio, exergy loss, and exergy efficiency of quince fruit dur-
ing the drying process with high accuracy of with R2 value of 0.9997, 0.9989,
0.9988,0.9986, and 0.9978, respectively
(ii) The ANFIS model was compared with the ANN model, and the results
obtained showed that the ANFIS model has higher accuracy compared to the
ANN model where the value of R2 was higher with a lower error value in the
ANFIS model
Rapeseed oil (i) The developed ANFIS could predict the different outputs of rapeseed oil pro- Farzaneh et al. [152, 153]
cess by oil extraction and cooking at industrial scale, and the model achieved a
high correlation coefficient which is around 0.99
Salmonella enteritidis (i) Prediction of the inactivation of Salmonella enteritidis by ultrasound was able Soleimanzadeh et al. [191]
to be done by the developed ANFIS model with a good accuracy where the
correlation coefficient obtained was 0.974
(ii) This study was known to be important in the food industry as the bacteria
can cause food poisoning if proper detection is not being done
Taro (i) The optimization of extraction conditions of antioxidants from the taro flour Kumar & Sharma [192]
can be done by using the developed ANFIS model coupled with response
surface methodology
(ii) The prediction values obtained from the developed model were validated by
comparing with the experimental values, and the results were almost consistent
with prediction values from the developed model
Vegetables (can- (i) The developed model by using the ANFIS system was able to predict the Kaveh et al. [193]
taloupe, garlic, effective moisture diffusivity, specific energy consumption, moisture ratio,
potatoes) and drying rate of the vegetables with a high regression coefficient of 0.9990,
0.9917, 0.9974, and 0.9901, respectively, with minimum error value
(ii) Comparison between the ANFIS model and ANN model was carried out,
and the results showed that the ANFIS model possess a higher efficiency than
that of ANN model
Virgin olive oil (i) The ANFIS model was able to predict the quality of virgin olive oil samples Arabameri et al. [194]
with high accuracy where the coefficient determination obtained was greater
than 0.998
(ii) It was also able to visualize the effects of temperature, time, total polyphenol,
fatty acid profile, and tocopherol on the oxidative stability of virgin olive oil
Yam (i) The prediction of the yam moisture ratio during the drying process showed a Ojediran et al. [195]
high R2 value with 0.98226 by using the developed ANFIS model
Food Engineering Reviews (2022) 14:134–175 149

(PARC) and multivariate data analysis [102]. For example, soft computing techniques has been said as a valuable and
different types of chemical substances in the liquid phase important tool in the food industry. This is because the com-
samples can be segregated using E-tongue. About seven bination of these systems offers good advantages such as an
sensors of electronic instruments are equipped in E-tongue, accurate prediction in a fast manner can be achieved. Table 8
which enabled it to identify the organic and inorganic com- shows the combination of CVS and soft computing that has
pounds. A unique fingerprint is formed from the combina- been used in the food industry. Figure 8 shows the working
tion of all sensors that has a spectrum of reactions that differ principle of CVS. An example on the utilization of CVS for
from one another. The statistical software of E-tongue ena- the quality control is shown in Fig. 9 [113].
bles the recognition and the perception of the taste. E-tongue Near infrared spectroscopy (NIRS) is another technique in
comprises three elements specifically the sample-dispensing the food industry as there is no usage of chemicals and results
chamber or automatic sample dispenser, an array of sen- can be obtained accurately as well as precisely within minutes
sors of different selectivity, and image recognition system or even continuously [114]. In addition, it is known to be non-
for data processing (Ekezie, 2015). Samples in liquid forms destructive, cost effective, quick, and straightforward which
could be analyzed directly without any preparation while makes it a good alternative for the traditional techniques which
the samples in solid forms have to undergo preliminary dis- are expensive and labor intensive and consumes a lot of time
solution before measurement is carried out. The process of [115]. The chemical-free method by NIRS makes it suitable to
E-tongue system is shown in Fig. 7 below. The ability to be used as a sustainable alternative since it will not endanger
sense any taste like a human olfactory system makes it one the environment or the human health. It has a wide range of
of the important devices in the food industry, especially for quantitative and qualitative analysis of gases, materials, slur-
quality control and assurance of food and beverages [103]. ries, powders, and solid materials. Furthermore, samples are
In addition, E-tongue has been used to identify the aging of not required to be grounded when light passes through it and
flavor in beverages [104], identify the umami taste in the certain features or characteristics that are unique to the class
mushrooms [105], and assess the bitterness of drinks or dis- of the sample are revealed by the spectra of the light. Com-
solved compounds [102]. Jiang et al. performed a summa- plex physical and chemical information on the vibrational of
rized review on the application of e-nose in the sensory and molecular bonds such as C–H, N–H, and O–H groups and
safety index detection of foods [106]. Moreover, the demand N–O, C–N, C–O, and C–C groups in organic materials can be
of E-tongue in the food industry market has risen due to the provided by the spectra which can be recorded in reflection,
awareness on delivering safe and higher-quality products. interactance, or in transmission modes [114].
The details of recent applications of E-tongue in the food The basic working principle for NIRS is shown in Fig. 10.
industry are shown in Table 7. Recently, NIRS has become an interest in food industries to
The computer vision system (CVS) is a branch of AI that inspect food quality, controlling the objective of the study
combines the image processing and pattern recognition tech- and evaluating the safety of the food [114, 116–119]. Several
niques. It is a non-destructive method that allows the exami- researchers have applied the NIRS in food to obtain its proper-
nation and extraction of image’s features to facilitate and ties for multiple reasons including determining the fatty acid
design the classification pattern [107]. It is also recognized profile of the milk as well as fat groups in goat milk [120].
as a useful tool in extracting the external feature measure- Apart from that, it is able to aid in the prediction of salted meat
ment such as the size, shape, color, and defects. In general, it composition at different temperatures [121] and in the pre-
comprised a digital camera, a lighting system, and a software diction of sodium contents in processed meat products [122].
to process the images and carry out the analysis [108]. The The detection and grading of the wooden breast syndrome in
system can be divided into two types which are 2D and 3D chicken fillet in the process line was also able to be performed
versions. Its usage is not restricted to various applications by using the NIRS technique [123]. Not only that, it is proven
in food industries such as evaluating the stages of ripeness to be efficient in determining the maturity of the avocado based
in apples [107], predicting the color attributes of the pork on their oil content [124], predicting the acrylamide content
loin [109], detecting the roasting degree of the coffee [110], in French-fried potato and in the potato flour model system
evaluating the quality of table grapes [111], and detecting [125], and determining the composition of fatty acid in lamb
the defects in the pork [112]. The combination of CVS with [126]. There has been a review conducted on the application

Fig. 6  E-nose system

Array of Intelligent
Samples Compounds Signals Processed Data Results
Sensors Algorithm
Table 6  Application of E-nose with AI in food industries
150

Application Objectives AI technique Outcomes/impacts References

Beef To classify the beef samples Adaptive FL system (AFLS); (i) The accuracy of the AFLS model was very Kodogiannis & Alshejari [196]
ANFIS good which was 94.28% for overall correct
classification which shows that it is able to
tackle complex, non-linear problems like
meat spoilage
(ii) By using the ANFIS method, the results
were satisfactory but were obtained with
high computational cost
(iii) E-nose with appropriate machine learn-
ing tools proved that it is useful to monitor
the spoilage of the meat during aerobic
stage at various temperatures
Cocoa To figure out and classify the fermentation DT, boosted tree RF, ANN, KNN, naïve (i) The ANN and Boosted tree algorithm Tan et al. [197]
time of the cocoa beans Baiyes (NB) manage to obtain an acceptable classifica-
tion rate while the fermentation time was
not able to be determined by the KNN and
NB algorithm
Coffee beans To forecast the level of acidity in fresh ANN (i) The model was able to predict the acidity Thazin et al. [198]
roasted beans level values with an accuracy around 95%
based on human sourness level scores
Chicken meat To classify the fresh and freeze-thawed Fuzzy K-nearest neighbors algorithm (FK- (i) The FK-NN algorithm showed a high Mirzaee-Ghaleh et al. [199]
chicken meat NN) performance, and it can be used in e-nose to
identify and classify the fresh and frozen-
thawed chicken meat
Cow ghee To detect the adulteration of the margarine in ANN (i) The ANN model was able to analyze the Ayari et al. [200]
cow ghee data obtained from the e-nose with high
accuracy
Edible oil To detect the adulteration in oxidized and ANN (i) The developed ANN model with e-nose Karami et al. [201]
non-oxidized edible oil was able to detect the adulteration in the
edible oil with high accuracy
(ii) The classification of the system was com-
pared with other methods and it was given
that ANN had the highest classification rate
with 97.3%
Fish To identify and classify the fish spoilage ANN, PCA (i) The developed model using the PCA and Vajdi et al. [202]
ANN was able to classify the fish according
to their spoilage group with an accuracy of
96.87%
Food Engineering Reviews (2022) 14:134–175
Table 6  (continued)
Application Objectives AI technique Outcomes/impacts References

Fruits To apply Kernel extreme learning machines Kernel extreme machine learning (i) The proposed system performed signifi- Uçar & Özalp [203]
in E-nose for recognition and perception of cantly good in odor recognition
fruit odors (ii) It achieved higher testing accuracy and
smallest value of training time and testing
time compared to other systems that were
used for the comparison for recognizing the
fruit odors
Fruits To apply artificial bee colony (ABC) algo- ANN & artificial bee colony (i) ABC-ANN is more successful in odor Adak & Yumusak [204]
rithm in ANN to classify data from elec- classification of e-nose data and has higher
tronic nose and evaluate its performance performance compared to the backpropaga-
Food Engineering Reviews (2022) 14:134–175

tion algorithm that was used

(ii) The system was able to classify four
different aroma which are the aroma of the
lemon, cherry, strawberry, and melon
Fruit juice To figure out the amount of the food additives RF, SVM, ELM, PLSR (i) The additives which are benzoic acid and Qiu & Wang [205]
in the fruit juice chitosan in the juice could be predicted
with accurately by using the ELM and RF
methods with R2 value of 0.92 and 0.91,
respectively
(ii) ELM and RF method has a higher
accuracy in predicting the additives when
compared to ELM and PLSR where the R2
value obtained by ELM and PLSR are 0.72
and 0.51, respectively
Honey To determine the best classifier in forecasting ANN, SVM (i) Both the ANN and SVM model combined Faal et al. [206]
the physicochemical properties of Iranian with the e-nose showed a good prediction
ziziphus honey samples rate in determining the physicochemical
properties of the honey sample
(ii) The developed model using ANN method
shower a higher accuracy in the prediction
compared to that of with SVM method
Lemon To characterize and predict the quality of ELM, RF, SVM (i) Prediction using the ELM method Guo et al. [207]
various lemon slices obtained the highest accuracy of 0.959
followed by RF and SVM with 0.935 and
0.922 respectively which indicates that
ELM integrated with the E-nose is the best
classification model
Pear To enhance the food quality by optimize the FL (i) The drying time was able to be shortened Li et al. [208]
drying process of balsam pear to retain the aroma and enhance the product
quality
(ii) Designation of industrialized control
method was established to simplify the
control and has good drying effects
151
Table 6  (continued)
152

Application Objectives AI technique Outcomes/impacts References

Pork meat To differentiate between the fresh and frozen- BPANN (i) Combined ANN with E-nose was able to Górska-Horczyczak et al.
thawed meat distinguish three types of meat which are [209]
loin, neck and ham
(ii) The model was able to differentiate the
fresh meat from spoiled meat and frozen
meat with overall sensitivity of 85.1% and
97.5% specificity
Rice grains To detect the Sitophilus oryzae infestation in FL ARTMAP (i) E-nose with the application of FL ART- Srivastava et al. [210]
stored rice grains PCA, MLR MAP is useful when the data are exhaustive
and deductions about analysis are needed
to be done
(ii) The model was able to classify the grains
of infested, non-infested, required treatment
and others
(iii) The hybrid system is beneficial in the
food and grain industry where early detec-
tion of infestation in grain can be done to
minimize the post-harvest losses
Rice grains To classify the Rhyzopertha. Dominica ANN (i) BPANN with e-nose system gave the Srivastava et al. [211]
infested rice grains and recognized them highest R2 value which is 0.98 compared
to other methods that were used with the
e-nose for the classification process and it
has the highest accuracy compared to the
rest
(ii) The system was able to predict the
infested rice grains for different days
Shelled peanuts To assess the storage quality of shelled FL (i) The FL analysis able to screen and rank Raigar et al. [212]
peanuts the e-nose sensors and the discard time
for the shelled peanuts was able to be
determined
(ii) The storage time prediction of shelled
peanuts using sensors closely matched with
the conventional methods
(iii) It can be an eco-friendly alternative as it
is rapid and not destructive
Salmon To classify the level of freshness of the PCA, CNN, SVM (i) The developed system was able to detect, Feng et al. [213]
salmon samples cluster, and classify the salmon accord-
ing to their freshness level, and the overall
accuracy obtained by the CNN-SVM model
was greater than 90%
Saffron To detect the adulteration of saffron samples ANN (i) The classification of original and adulter- Heidarbeigi et al. [214]
ated saffron was successfully done by using
e-nose with the pattern recognition method
with 100 and 86.87% accuracy, respectively
Food Engineering Reviews (2022) 14:134–175
Table 6  (continued)
Application Objectives AI technique Outcomes/impacts References

Saffron To identify different aromas of Iranian saffron ES, PLS, ANN (i) E-nose coupled with ANN successfully Kiani et al. [215]
obtained 100% for the classification of saf-
fron samples
Spinach To detect the postharvest freshness of spinach ANN (i) Combination of E-nose with the BPNN Huang et al. [216, 231]
method proves to be effective for a fast and
non-destructive way to detect the spinach
freshness with an accuracy of 93.75% for
the classification accuracy
Strawberry juice To classify and carry out analysis by using ANN, (i) E-nose could differentiate each treatment Qiu et al. [217]
E-nose with neural networks and other ML type and be able to predict the quality
Food Engineering Reviews (2022) 14:134–175

learning methods as well to evaluate the parameters

performance of the system (ii) ELM network was able to classify the
strawberry juice and is much faster com-
pared to other modelling networks that were
used
(iii) ELM also showed a better performance
compared to other modelling techniques
that were being used
Sunflower oil To determine the frying disposal point of SO FL (i) E-nose combined with FL was able to Upadhyay et al. [218]
(SO) assess the frying disposal time of fried SO
blend
(ii) The combination of e-nose with FL has
potential for other fried product platforms
Wheat grain To determine the granary weevil infestation FL (i) The most responsive sensors and specific Mishra et al. [219]
in stored wheat grains VOCs generated by insect-infected wheat
grains were able to be screened out by the
e-noses sensor associated with the fuzzy
logic analysis
(ii) E-nose was proven to be a potential
method for accurate and rapid in monitor-
ing the infestation in stored wheat grain.
It is also a reliable method for industries
to determine the quality of the product
throughout the storage period
153
154 Food Engineering Reviews (2022) 14:134–175

Fig. 7  E-tongue system

of the ANN combined with the near-infrared spectroscopy for in the food industry. However, decent data are required for
the detection and authenticity of the food [127]. The ability the model to learn in order to perform well. In addition to
of the NIRS system in detecting the physical and chemical that, this model is useful for solving analytical mathematical
properties coupled with soft computing techniques such as models in the food industry such as studies involving mass
ANN, FL, and ML allows the classification and prediction of and heat transfer coefficients. ANFIS is recommended to be
the samples to be performed rapidly and accurately. Table 9 used when complex systems where time-varying processes
shows the application of NIRS coupled with AI techniques in or complex functional relationships and multivariable are
the food industry. involved. Apart from that, it can be used in descriptive sen-
sory evaluation.
These AI algorithms can be combined with other sen-
Summary on the Application of AI sors such as the electronic nose, electronic tongue, computer
in the Food Industry vision system, and near infrared spectroscopy to glean the
data from the samples. Both the E-nose and E-tongue have
From the review so far, it can be shown that AI has been shown to enhance the quality characteristics in comparison
used for various reasons in food industries such as for detec- to the traditional detection approach [128]. E-nose can be
tion, safety, prediction, control tool, quality analysis, and used to sense the odors or gases while the E-tongue can be
classification purposes. Ranking of sensory attributes in the applied for the identification of the organic and inorganic
foods can be done easily by using the FL model. Not only compounds. Studies involving the examination and drawing
that, fuzzy logic can be used for classification, control, and out the features of the samples like shape, color, defects,
non-linear food modeling in the food industry. ES is widely and size can be carried out by using the CVS sensors. NIRS
used in the food industry for decision-making process. On can be utilized to determine the properties or contents in
the other hand, ANN model is applied widely in the food the samples. The data obtained from these sensors is then
industry for prediction, classification, and control task as merged with the AI algorithms and utilizing their computing
well as for food processing and technology. The supervised strengths to accomplish the desired studies.
ANN method has the ability to learn from examples which
allows for the prediction process to be done accurately.
Meanwhile, the unsupervised method of ANN is found to Advantages and Disadvantages of AI
be more common for the classification task. Another method
that has been utilized for the prediction and classification AI has been used widely in the industry as it offers a lot
of the food samples is by using the machine learning (ML) of advantages compared to the traditional method. All
method. ML can be used in solving complicated tasks which the algorithms are known to be accurate and reliable,
involves a huge amount of data and variables but does not but careful selection should be made by considering the
have pre-existing equations or formula. This method is advantages and limitations of the algorithms. The differ-
known to be useful when the rules are too complex and ent algorithms have their own strengths and weakness,
constantly changing or when the data keep changing and hence choosing them for a particular application in the
require adaptation. Furthermore, the adaptive neuro fuzzy food industry needs to be looked on a case-to-case basis.
inference system (ANFIS) is another hybrid AI method that The guideline to choose the most appropriate method is
can be used to solve sophisticated and practical problems given in the next section. The benefits and constraints
Table 7  Application of E-tongue with AI in food industries
Application Objectives AI technique Outcomes/impacts References

Ham To monitor the salt processing of hams salted differ- Simplified fuzzy ARTMAP neural network (i) The data obtained from the e-tongue was able to Gil-Sánchez et al.
ently with different formulations be analyzed and classified using ANN [220]
(ii) The data was classified using two processing
variables which are the processing time and salt
formulation
(iii) Optimum parameters value for SFAM neural
networks were drawn out to be used in the micro-
controller device
Honey To differentiate different types of honey according Fuzzy ARTMAP neural network (FAM) (i) The proposed E-tongue system was able to dif- Marisol et al. [221]
to their antioxidant level ferentiate different types of honey as well as their
total antioxidant capacity level
Food Engineering Reviews (2022) 14:134–175

(ii) The ANN fuzzy art map type analysis had a high
classification success rate of 100% which indicates
that it is a good measurement system
Liquor To classify different types of Chinese liquor flavor SVM & fuzzy evaluation (i) E-nose with the SVM system was able to classify Jingjing et al. [222]
using e-tongue with fuzzy evaluation and predic- four different flavors of liquor with an accuracy
tion by SVM of 100%
(ii) The developed system is able to discriminate
the samples accurately and the output evaluation
language in line with the human perception
Milk To detect the adulteration of raw milk SVM (i) The developed model was able to determine the Tohidi et al. [223]
adulteration in the samples with a high accuracy
values which are all greater than 87% for different
types of adulterants in the milk
Peanut meal To assess the taste attributes of peanut and compare ANN, partial least square (PLS) (i) Good stability and repeatability with respect to Wang et al. [224, 225]
the predictive abilities of the methods used the measured signals were exhibited by the sen-
sors in the E-tongue
(ii) Different concentrations with the same taste
(five types of taste) were able to be discriminated
by the E-tongue
(iii) RBFNN has a better prediction ability with
lower error and higher correlation coefficients than
those of the PLS method
Pineapple To classify the pineapples according to their sweet- SVM, KNN, ANN, RF (i) Different machine learning algorithms were Hasan et al. [226]
ness level and determine the best algorithm employed in determining the sweetness of the
pineapple, and the best algorithm obtained was
the KNN method where it achieved an accuracy
of 0.820
(ii) The developed model will be beneficial in
industry when the selection of pineapples in large
quantities is required
155
Table 7  (continued)
156

Application Objectives AI technique Outcomes/impacts References

Rice To discriminate and predict the solid foods as well ANN (i) RBFNN was able to distinguish different types of Wang et al. [224, 225]
as to provide an assessment tool for food indus- rice with 95% accuracy in classification
tries (ii) Voltametric E-tongue is useful for the qualitative
analysis for rice
Rice To develop similarity analysis combined with artifi- ANN (i) SA-ANN in E-tongue was able to predict the rice Lu et al. [227]
cial neural networks (SA-ANN) in e-tongue for the sensory quality and carry out systematic analysis
prediction of rice sensory quality (ii) Comparison was carried out between PCA-ANN
and SA-ANN, and it was found that SA-ANN has
better precision and accuracy compared to PCA-
ANN
(iii) SA-ANN is a less-labor intensive, quicker
method and has potential for rapid and big scale
prediction of rice sensory property
Sugarcane To characterize and apply voltametric e-tongue for ANN (i) Multilayer ANN with wavelet information was De Sá et al. [228]
the analysis of glucose from the sugarcane able to process complex responses from the
E-tongue
(ii) The proposed model is suitable to be used for
hydrolyzed samples from sugarcane bases
Tangerine peel To classify tangerine peel of different ages BPNN, ELM (i) The model was able to classify the tangerine peel Shi et al. [229, 230]
samples of different ages
(ii) Comparison was done for few linear models and
non-linear models, and it was obtained that non-
linear models exhibited better performance than
linear models
(iii) ELM was the best for the classification of the
samples with high accuracy followed by BPNN
Teas To distinguish different types of teas ANN, SVM (i) Different types of teas were able to be distin- Huang et al. [216,
guished by using the developed system and the 231]
compositions of the tea also could be identified
Tilapia fillets To predict the changes in freshness of tilapia fillets ANN-PCA (i) E-tongue is able to distinguish the extracts of Shi et al. [229, 230]
at different temperatures using the combined tilapia fillets stored at different days and different
techniques temperatures
(ii) The model set up is able to predict the freshness
of tilapia fillets stored at different temperatures
ranging from 0 to 10 °C
Food Engineering Reviews (2022) 14:134–175
Table 8  Application of CVS with AI in food industries
Application Objectives AI technique Important outcomes References

Apple To sort the defective and normal apples CNN, SVM (i) The developed CNN with the CVS model was Fan et al. [232]
able to classify the apples with a high accuracy
rate of 96.5%, and it was proven to be more
effective than the conventional image processing
method which was combined with the SVM clas-
sifier where the accuracy rate was 87.1%
Apple slices To study the drying effects on the changing color of ANN (i) CVS was able to track the color changes during Nadian et al. [233]
the apple slices the drying process, and the combination with the
ANN was able to estimate the quality of the apple
during the drying process
ii) The developed model achieved R2 values of
Food Engineering Reviews (2022) 14:134–175

greater than 0.92 for all the analyses

Banana To classify the banana according to its ripeness ANN, SVM, KNN, DT (i) The ANN-based model system has a higher clas- Mazen & Nashat [234]
sification rate compared to the other algorithms
with the highest overall recognition rate of 97.75%
Barley flour To predict the barley flour based on the improvised SVM, KNN, DT, RF (i) The developed model of CVS with different Lopes et al. [235]
method learning algorithms was improvised by using the
spatial pyramid partition ensemble method for
the classification of the barley flour where the
accuracy achieved was 75% (KNN), 95% (SVM &
RF), and 100% (DT)
Beer To forecast the beer acceptability based on different ANN, ML (i) Seventeen ML algorithms were used to find the Gonzalez Viejo et al. [236]
sensory parameters best model with a good performance was carried
out, and the results showed that Bayesian regu-
larization had the best accuracy where the R value
obtained was 0.92
(ii) The combination of RoboBEER, CVS, and
ANN algorithms allowed to determine the beer
making based on its acceptability of customers
and its quality
Bell pepper To describe the ripeness level of bell pepper auto- ANN, FL (i) An artificial vision system was able to be devel- Villaseñor-Aguilar et al. [237]
matically oped by using the CVS and ANN/FL in predicting
the maturity of the bell pepper
(ii) The model using RBF-ANN has a higher clas-
sification accuracy compared to FL where the
maximum accuracy obtained by both models are
100% and 88%, respectively
Cape gooseberry To classify the ripeness of cape gooseberry ANN, SVM, DT, KNN (i) All the models were able to classify the ripeness Castro et al. [238]
of the cape gooseberry with a high accuracy
where the accuracy obtained by all the models
was greater than 86% using different color spaces,
which indicates that it is a good classifier system
157
Table 8  (continued)
158

Application Objectives AI technique Important outcomes References

Coffee beans To develop a system that can classify the coffee ANN (i) Green coffee beans were able to be analyzed and De Oliveira et al. [108]
beans classified by using the developed system
(ii) ANN were used in this system as a color space
transformation model where the transformed value
was later used for the classification purposes
Dry beans To classify different types of seed from the produc- ANN, KNN, DT, SVM (i) The classification of the beans was able to be Koklu & Ozkan [239]
tion done by all the ML algorithms with SVM which
achieved the highest overall classification rates
of 93.13% followed by the DT, ANN, and KNN
where the classification rates were 92.52%,
91.73%, and 87.92%, respectively
Eggs To predict the volume of eggs ANN (i) The volume of the eggs was predicted by a Siswantoro et al. [240]
developed system with a good linear coefficient
of 0.9738 with the actual volume and relative
absolute error of 2.2078% which indicates that the
developed system is an efficient model
Figs To classify fig fruits based on its visual features Decision tree-FL (i) Comparison was carried out between three dif- Khodaei & Behroozi-khazaei [241]
ferent types of decision tree for the data obtained
from CVS, and it was shown the REP decision
tree had the highest value of R and lower RMSE
values and hence was selected to be implemented
in the fuzzy system
(ii) The developed system was able to classify the
fig fruits into five qualitative grades with a high
accuracy
Fish To identify the freshness of the fish ANN, PCA (i) The developed system was able to classify the Huang et al. [242]
freshness of fish with a success rate of 94.17% in
the training set and 90.00% in the prediction set
Gluten-free cake To develop a system for quality control of celiac- FL (i) The developed system was able to study the tex- Rezagholi & Hesarinejad [243]
friendly products ture of the cake when different amounts of materi-
als were added to it, and the optimal ingredients
value suitable for the gluten-free cake were able to
be determined
Lime To predict the weight of Indian lime fruits ANFIS (i) Different clustering methods were fused with Phate et al. [244]
the ANFIS model to improve the accuracy in the
classification system, and it was found that the
fuzzy C-means clustering (FCM) was the best in
predicting the weight of the sweet lime
(ii) The developed system was able to predict the
weight of the Indian sweet lime fruits accurately
Food Engineering Reviews (2022) 14:134–175
Table 8  (continued)
Application Objectives AI technique Important outcomes References

Mango To estimate the fruit mass of the mango ANN (i) The data obtained from the CVS was used as the Utai et al. [245]
input parameters in the developed ANN model,
and the model was able to estimate the fruit mass
successfully
(ii) The developed system has the highest success
rate of 97% and the efficiency coefficient of 0.99
by applying two input parameters or three input
parameters
Mushrooms To determine the appearance quality of mushrooms ANN, FL (i) The accuracy obtained by the image processing Nadim et al. [246]
system was 95.6%
(ii) The artificial neural network was able to deter-
Food Engineering Reviews (2022) 14:134–175

mine the weight of the mushrooms, and the fuzzy

logic used the data from the CVS and was able to
determine the quality of the mushroom
Passion fruits To classify the passion fruits based on their ripeness Multi-class SVM (MCSVM) The developed system was able to classify the Sidehabi et al. [247]
level ripeness level passion fruits with an accuracy of
93.3% within 0.94128 s
Pork loin To assess the quality of the pork loin according to SVM (i) The model was able to predict the quality of the Sun et al. [248]
the industry demand pork loin based on their color and quality attrib-
utes based on the industries’ demand
Potatoes To develop a grading system for potatoes FL (i) Combination of CVS and fuzzy logic allows Bhagat & Markande [249]
faster grading of the potatoes as well as reduces
the cost required for the manual grading
Rice To carry out the qualitative grading of milled rice FL (i) The study was able to conclude that the devel- Zareiforoush et al. [250]
oped hybrid system can be used in the processing
industry for automatic grading of milled rice
(ii) Comparison was done between the developed
system and experts’ judgment, and around 89.80%
overall confidence was obtained
(iii) The fuzzy system has obtained 89.83% total
sensitivity and 97.45% specificity for the quality
grading of milled rice
Rice To control the performance of rice whitening FL (i) The developed automatic control system had an Zareiforoush et al. [55]
machines average of 31.3% higher performance speed than
that of a normal human operator, and there was an
improvement in the quality of the output based on
the decision made by the system
(ii) The setup flexibility of the system allows altera-
tion to be done according to the preference of each
rice mill operator
159
 160 Food Engineering Reviews (2022) 14:134–175

that each of the algorithm exhibits are explained briefly

in Table 10.
Bakhshipour et al. [251]

Guidelines on Choosing the Appropriate AI

Method

Garcia et al. [253]

Wan et al. [252]
References

Selecting the appropriate algorithm is important when

developing the AI model as it can aid the user to attain an
accurate, rapid, and cost-saving results. Therefore, a guide-
line given in Fig. 3 is an important asset prior to achieving
(ii) The research successfully showed that DT-based

ligent classification of Iranian green tea and black

fuzzy systems can be applied for automated intel-
venient compared to the J48 tree for developing a

best performances in a case study. The primary step in the

accuracy of 99.31% and 1.2% of standard devia-

(i) The system can classify the two different types

detected by using the developed system with an

of seeds that look similar which are cauliflower

(i) REP decision tree was shown to be more con-

selection process is that users should define and finalize the

(i) The maturity of the tomato was able to be

objective of using AI in their research or implementation.

Prediction, classification, quality control, detection of adul-
terants, and estimation are among the common objectives of
seed and Chinese cabbage seed

AI applications in the food industries. Next, decision should

be made whether sensors such as E-tongue, E-nose, CVS,
and NIRS are required to collect the sampling data or not for
fuzzy classifier system

collecting the data from the samples. Normally, integration

Important outcomes

with those sensors is conducted to obtain the parameters

and characteristics of the samples to be included in the AI
algorithms for sample testing purposes. Upon deciding the
necessity of the sensors, users should compare and choose
tion
tea

the fitting algorithm with respect to their study. Among

the most common AI algorithms that have been employed
include the FL, ANN, ANFIS, and ML methods. ANFIS
has shown to have a higher accuracy, but the complexity
of the model makes it less favorable compared to the other
algorithms. It is advisable for the users to determine the
AI technique

complexity of the research in selecting the most appropriate

algorithm for their studies. Once the selection of the algo-
DT-FL

ANN

rithm has been confirmed, the data available are integrated

FL

with the AI algorithms. Finally, the testing and validation

based on R2 and MSE are done to analyze the performance
of the established model. The AI model has been created
successfully once the validation is accepted; otherwise, users
To classify the Iranian green and black tea

To detect the maturity of the fresh tomato

should return to the previous step and reselect the algorithm.

Figure 11 shows the guideline in choosing and development
of the AI model in food industry application.
To classify the vegetable seeds

Trends on the Application of AI in the Food

Industry in the Future
Objectives

The overall trend on the application of AI in the food indus-

try is shown in Fig. 12. From the studies within the past few
years, the usage of the AI methods has been observed to
Table 8  (continued)

increase from 2015 to 2020 and is predicted to rise for the

Vegetable seeds

next 10 years based on the current trends. Among the rising

Application

factors for the application of AI in the food industry is the

Tomato

introduction of Industrial Revolution 4.0 (IR 4.0). The merg-

Tea

ing of technologies or intelligent systems into conventional

Food Engineering Reviews (2022) 14:134–175 161

Fig. 8  Working Principle of
CVS Interpreting
Input Sensing Device Output
Device

industry is what is known as IR 4.0 and can also be called for the upcoming years due to increasing in food demand
smart factory [129, 130]. AI which is categorized under the and the concern on the safety of the foods which are being
IR 4.0 technologies focuses on the development of intel- produced.
ligent machines that functions like the humans [131]. IR The comparison between the AI integration with and
4.0 makes a great impact in the product recalls due to the without sensors for real-time monitoring in the food industry
inspections or complains in the food industries. The imple- is displayed in Fig. 13. Integration with external sensors has
mentation of the AI integrated in the sensors able to detect a higher percentage compared to those without the integra-
the errors during the manufacturing process and rectify the tion of the sensors in the food industries. The purpose of
problems efficiently. Apart from that, IR 4.0 also plays a big external sensors was to obtain the data from the samples
role in the human behavior as consumers in the twenty-first which are then employed into the AI algorithms to carry
century often discover information regarding the foods in out various tasks such as classification, prediction, quality
the internet. The rising concerns on the food quality allow control, and others that have been stated earlier. However,
more usage of AI as they are able to enhance the quality of the data collection for the year 2017 showed that the percent-
the food and aids during the production process. The highest age for the AI without the external sensors is greater than
amount of application of AI in the food industry was seen in that with integration with the sensors. This is due to the high
the year 2020 as more researchers are carrying out studies amount of research which was conducted without using the
using the AI method, and it is believed to continue rising external sensors which are listed in this paper. Based on the

Fig. 9  CVS-based quality con-

trol process

Fig. 10  Basic working principle

of CVS
Table 9  Application of NIRS with AI in food industries
162

Application Objectives AI technique Important outcomes References

Cheese To determine the characteristics of controlled- PCA-ANN (i) Sensory attributes of cheese have been deter- Curto et al. [254]
processing cheese mined with sufficient evidence on well-defined
population which was known to be cost effective
and useful method for quality control
Chicken meat To classify chicken meat with respect to their qual- Decision tree (DTree) (i) The assessment of the poultry meat was able Barbon et al. [255]
ity grades to be done by using the spectral analysis and the
classification using the DTree models (REPTree)
had a better performance compared to the sup-
port vector machine and decision table model
that were carried out for comparison purposes
Chocolate To develop a model for the quality assessment of ANN (i) NIRS was used to obtain the fingerprint of the Gunaratne et al. [256]
chocolate chocolate using the five taste senses and fused
with ANN for the prediction of chocolate quality
based on chemical, physical, and sensory proper-
ties and accurate prediction was achieved by the
developed model
Civet coffee To detect any alteration in the civet coffee ANN, SVM, KNN (i) The combination of trained ANN and NIRS Arboleda [257]
method was able to discriminate the civet coffee
from non-civet coffee successfully with an accu-
racy in the range of 95–100%
(ii) The performance of classification using ANN
was compared with other learning algorithms
such as decision trees, discriminant analysis,
SVM, KNN, and ensemble classifiers, and it was
proven that ANN has better accuracy compared
to the others
Gelatin To detect the adulteration in edible gelatin ANN, SVM (i) NIRS fused with both methods which were Zhang et al. [258]
done separately was able to detect the adultera-
tion in edible gelatin
(ii) SVM model showed as the best recognition
methods among the other models that were
used with an accuracy of 100% meanwhile the
recognition rate obtained by the ANN model was
97.44%
Eggs To determine the freshness of the egg RBFNN, PCA (i) The system was able to classify the egg fresh- Aboonajmi et al. [259]
ness, and the model is appropriate to be imple-
mented for the rough screening of the eggs
Food powders To classify different types of food powders SVM (i) The system was able to classify the food pow- Mohamed et al. [260]
ders varying of whole wheat flour, organic wheat
flour, tapioca starch, corn starch, and rice flour
with a very high accuracy rate of 100%
Food Engineering Reviews (2022) 14:134–175
Table 9  (continued)
Application Objectives AI technique Important outcomes References

Keemun black tea To develop a model for the discrimination of differ- ANN, least square SVM (i) Three different methods which are ANN and Ren et al. [261]
ent grades of keemun black tea in China (LSSVM), Random Forest LSSVM were used to develop the model and all
the methods were able to distinguish the different
grades of keemun tea
(ii) It was identified that the LSSVM method has
a higher performance and predictivity compared
to the rest
Meat To create a system for the detection of meat spoil- AFLS (i) The AFLS model was able to classify the meat Alshejari & Kodogiannis
age into three classes which are fresh, semi-fresh, [262]
and spoiled by using the data provided by the
FTIR spectrometer
Food Engineering Reviews (2022) 14:134–175

(ii) The model achieved a high percentage with a

value of 95.94% of correct classification overall
which indicates that it can be an effective tool for
the detection of meat spoilage
Olive oil To detect the adulteration in olive oils SVM (i) The fusion of NIRS and Raman spectral data Xu et al. [263]
could identify the adulterated olive oils effec-
tively and the SVM model was able to predict the
dopant contents in olive oil accurately
Pears To determine the soluble solid contents in pear ELM (i) The proposed successive projection algorithm Lu et al. [264]
and extreme learning machine (SPA-ELM) was
able to predict the contents better than the con-
ventional PCA-ELM method
Rice To classify the rice according to the compositions RF, PCA, LDA, PLS (i) All the ML techniques were able to identify and Rizwana & Hazarika [265]
and processing parameters classify the rice based on its composition (amyl-
ose-based, glycemic index) and the hydrothermal
treatment severity with a good performance
Wheat flour To predict the wheat flour quality using the NIRS Multitarget coupled with SVM, RF (i) The use of multitarget over partial least squares Barbon Junior et al. [266]
coupled with machine learning algorithm offers
more advantage for the parameter prediction
from NIRS
(ii) The prediction using random forest overcomes
the performance of SVM
White asparagus To predict the origin of the asparagus and distin- SVM (i) The linear SVM could predict the country of Richter et al. [267]
guish the German from imported products origin of white asparagus with an accuracy of
89% and also was able to distinguish the German
and non-German products
163
164 Food Engineering Reviews (2022) 14:134–175

Table 10  Advantages and limitations of AI algorithms

AI algorithm Advantages Limitations

ES (i) Reliable and understandable (i) The construction and designing of the ES are expensive and
(ii) Highly responsive rare as it requires expert engineers
(iii) High performance (ii) Vocabulary utilized by the experts is limited and often is
(iv) Error rate is lower than human errors difficult to be understood
(v) Better use of production capacities
FL (i) Imprecise, incomplete, and uncertain information can be (i) Generalization is not possible as it can only deduce the given
solved rules
(ii) Simpler and direct results can be obtained (ii) Sometimes requires the knowledge of an expert in creating
(iii) Accountable, noise tolerant, and robust to disturbances the rules
(iv) Faster interpretation than ANN and support vector
machine method
(v) The knowledge base can be extended easily with the exten-
sion of the rules
(vi) Saves costs and time
(vii) Can improve the quality and safety of the products
ANN (i) Able to model complex functions or problems accurately (i) The performance of the model is hard to be explained com-
and easily pared to the others as it appears as a black box model
(ii) Accurate, robust to disturbances, and noise tolerant (ii) Requires more time compared to the FL as the suitable num-
(iii) Has the ability to learn from the patterns or examples ber of layers should be determined
(iv) Has the generalization ability (iii) Need sufficient and reliable data
(v) Affordable, noise tolerant, easy, and flexible method
(vi) Solving non-linear problems are more appropriate by using
this method
(vii) Useful as prediction, classification, and control tool
ANFIS (i) Able to merge details from various resources (i) The data available should be reliable to avoid any confusion
(ii) Noise tolerant, accurate, and effective method in solving or misinformation during the training process as it will affect
complex problems the final results
(iii) It has a higher performance compared to ANN and FL
methods
(iv) Possesses the benefits from both ANN and FL method
(v) Classification and prediction tasks can be done more
conveniently
(vi) Able to save time and cost overall in general compared to
manual methods

evaluation carried out during this study, it was found that the main factors which are affecting the food industries.
a high amount of research was done on the integration of Apart from that, in the recent pandemic situation due to
CVS sensors with the AI methods. It is explainable as CVS the Covid-19 virus, it is predicted that more companies
sensors are able to provide important parameters such as the will opt for the usage of AI in their industries to cut down
shape, size, colors, and defects which are essential for the the costs and boost the performance of their company.
quality control in the food industry. However, the integra- There have been reports by some of the SMEs that their
tion of the system is mainly dependent on the objectives of earnings have dropped and some SMEs have claimed that
the researcher and the industrial players and the availability they could only survive for about 1 to 3 months. The high
of the data. demand of food and the tight standard operating proce-
In short, as the AI world is heading towards 2.0 [132], it dure in the companies during the pandemic situation will
can be predicted that the rise in the usage of AI in the food encourage the industry players to find an alternative to
industry is definite and inevitable because of the advan- their problems and AI will be one of them to ensure a
tages that they can offer such as saving in terms of time, smooth operation.
money, and energy as well as the accuracy in predicting
Food Engineering Reviews (2022) 14:134–175 165

Fig. 11  Flowchart for developing AI model

Fig. 12  Application of AI in the Application of AI in Food Industry Over the Years

food industry
50

40
Percentage (%)

30

20

10

0
2014 2016 2018 2020 2022 2024 2026 2028 2030
Year

Fig. 13  Comparison between Comparison between AI Integra on With and Without

integration of AI for real-time
monitoring in the food industry External Sensors for Real Time Monitoring in Food Industry
80
Percentage (%)

60
40
20
0
2015 2016 2017 2018 2019 2020
Year

With No Integra on With Integra on

166 Food Engineering Reviews (2022) 14:134–175

Conclusion and Future Outlook 4. Narvekar M, Fargose P (2015) Daily weather forecasting using
artificial neural network. International Journal of Computer
Applications 121(22):9–13. https://​doi.​org/​10.​5120/​21830-​5088
In conclusion, AI has been playing a major role in the food 5. Waltham M, Moodley D (2016) An analysis of artificial intel-
industry for various intents such as for modeling, prediction, ligence techniques in multiplayer online battle arena game envi-
control tool, food drying, sensory evaluation, quality con- ronments. ACM Int Conf Proceeding Ser 26–28-Sept. https://d​ oi.​
org/​10.​1145/​29874​91.​29875​13
trol, and solving complex problems in the food processing. 6. Iqbal J, Khan ZH, Khalid A (2017) Prospects of robotics in food
Apart from that, AI is able to enhance the business strate- industry. Food Science and Technology 37(2):159–165. https://​
gies due to its ability in conducting the sales prediction and doi.​org/​10.​1590/​1678-​457X.​14616
allowing the yield increment. AI is recognized widely due 7. Ge Z, Song Z, Ding SX, Huang B (2017) Data mining and analyt-
ics in the process industry: the role of machine learning. IEEE
to its simplicity, accuracy, and cost-saving method in the Access 5:20590–20616. https://​doi.​org/​10.​1109/​ACCESS.​2017.​
food industry. The applications of AI, its advantages, and 27568​72
limitations as well as the integration of the algorithms with 8. Allawi MF, Jaafar O, Ehteram M, Mohamad Hamzah F, El-Shafie A
different sensors such as E-nose and E-tongue in the food (2018) Synchronizing artificial intelligence models for operating the
dam and reservoir system. Water Resour Manage 32(10):3373–3389.
industry are critically summarized. Moreover, a guideline https://​doi.​org/​10.​1007/​s11269-​018-​1996-3
has been proposed as a step-by-step procedure in develop- 9. Kawakami E, Tabata J, Yanaihara N, Ishikawa T, Koseki K, Iida
ing the appropriate algorithm prior to using the AI model Y, Saito M, Komazaki H, Shapiro JS, Goto C, Akiyama Y, Saito
in the food industry–related field, all of which will aid and R, Saito M, Takano H, Yamada K, Okamoto A (2019) Appli-
cation of artificial intelligence for preoperative diagnostic and
encourage researchers and industrial players to venture into prognostic prediction in epithelial ovarian cancer based on blood
the current technology that has been proven to provide bet- biomarkers. Clin Cancer Res 25(10):3006–3015. https://​doi.​org/​
ter outcome. 10.​1158/​1078-​0432.​CCR-​18-​3378
10. Nor Muhammad NA, Abdul Jalal AA (2019) Artificial neural
network based ovarian cancer survivability prediction tool. DSx-
Conference. https://s​ chola​ r.g​ oogle.e​ s/s​ chola​ r?h​ l=e​ s&a​ s_s​ dt=0%​
Funding  The authors were supported by the Universiti Kebangsaan
2C5&q=​Funci​onali​dad+​Famil​iar+​en+​Alumn​os+​de+1%​C2%​
Malaysia under grant GUP-2019–012.
B0+y+​2%​C2%​B0+​grado+​de+​secun​daria+​de+​la+​insti​tuci%​
C3%​B3n+​educa​tiva+​parro​quial+%​E2%​80%​9CPeq​ue%​C3%​
B1a+​Bel%​C3%​A9n%​E2%​80%​9D+​en+​la+​comun​idad+​de+
​Peral​v illo%​2 C+​u bica​d a+​e n+​e l+​d istr ​i to+​d e+​C hanc​ay+-+​
Declarations  perio​do+​2018&​btnG=
11. Ramakrishna RR, Hamid ZA, Zaki WMDW, Huddin AB, Mathi-
Conflict of Interest  The authors declare no competing interests. alagan R (2020) Stem cell imaging through convolutional neural
networks: current issues and future directions in artificial intel-
Open Access  This article is licensed under a Creative Commons Attri- ligence technology PeerJ 8 https://​doi.​org/​10.​7717/​peerj.​10346
bution 4.0 International License, which permits use, sharing, adapta- 12. Mahadevappa J, Groß F, Delgado A (2017) Fuzzy logic based
tion, distribution and reproduction in any medium or format, as long process control strategy for effective sheeting of wheat dough
as you give appropriate credit to the original author(s) and the source, in small and medium-sized enterprises. J Food Eng 199:93–99.
provide a link to the Creative Commons licence, and indicate if changes https://​doi.​org/​10.​1016/j.​jfood​eng.​2016.​12.​013
were made. The images or other third party material in this article are 13. Elferink M, Schierhorn F (2016) Global demand for food is ris-
included in the article’s Creative Commons licence, unless indicated ing. Can we meet it? Harvard Business Review 7(4):1–7. https://​
otherwise in a credit line to the material. If material is not included in www.​resea​rchga​te.​net/​publi​cation/​30246​6629%​0A
the article’s Creative Commons licence and your intended use is not 14. Garver K (2018) 6 examples of artificial intelligence in the food
permitted by statutory regulation or exceeds the permitted use, you will industry. Retrieved from https://​foodi​ndust​r yexe​cutive.​com/6-​
need to obtain permission directly from the copyright holder. To view a examp​les-​of-​artif​i cial-​intel​ligen​ce-​in-​the-​food-​indus​try/.
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. 15. Sharma Sagar (2019) How artificial intelligence is revolution-
izing food processing business? Retrieved from: https://​towar​
dsdata​ scien​ ce.c​ om/h​ ow-a​ rtif​i cial-i​ ntell​ igenc​ e-i​ s-r​ evolu​ tioni​ zing-​
food-​proce​ssing-​busin​ess-​d2a64​40c03​Cite Refer​ence.​60
16. Utermohlen K (2019) 4 Applications of artificial intelligence
References in the food industry, Retrived from https://​heart​beat.​fritz.​ai/4-​
appli​catio​ns-​of-​artif​i cial-​intel​ligen​ce-​ai-​in-​the-​food-​indus​try-​
1. Krittanawong C, Zhang H, Wang Z, Aydar M, Kitai T (2017) e742d​7c029​48
Artificial Intelligence in Precision Cardiovascular Medicine 17. Funes E, Allouche Y, Beltrán G, Jiménez A (2015) A review:
69(21):2657–2664. https://​doi.​org/​10.​1016/j.​jacc.​2017.​03.​571 artificial neural networks as tool for control food industry pro-
2. Hamet P, Tremblay J (2017) Artificial intelligence in medicine. cess. Journal of Sensor Technology 05(01):28–43. https://​doi.​
Metabolism: Clinical and Experimental 69, S36–S40. https://d​ oi.​ org/​10.​4236/​jst.​2015.​51004
org/​10.​1016/j.​metab​ol.​2017.​01.​011 18. Correa DA, Montero Castillo PM, Martelo RJ (2018) Neural
3. Borana J, Jodhpur NU (2016) Applications of artificial intelligence networks in food industry. Contemp Eng Sci 11(37):1807–1826.
& associated technologies. Proceeding of International Conference https://​doi.​org/​10.​12988/​ces.​2018.​84141
on Emerging Technologies in Engineering, Biomedical, Manage- 19. Kondakci T, Zhou W (2017) Recent applications of advanced
ment and Science [ETEBMS-2016], March, 5–6. control techniques in food industry. Food Bioprocess Technol
10(3):522–542. https://​doi.​org/​10.​1007/​s11947-​016-​1831-x
Food Engineering Reviews (2022) 14:134–175 167

20. Wang J, Yue H, Zhou Z (2017) An improved traceability system 37. Corney D (2002) Food bytes: intelligent systems in the food
for food quality assurance and evaluation based on fuzzy clas- industry. British Food Journal 104(10):787–805. https://​doi.​org/​
sification and neural network. Food Control 79:363–370. https://​ 10.​1108/​00070​70021​04488​90
doi.​org/​10.​1016/j.​foodc​ont.​2017.​04.​013 38. Perrot N, Ioannou I, Allais I, Curt C, Hossenlopp J, Trystram G
21. Alizadeh-Sani M, Mohammadian E, Rhim JW, Jafari SM (2020) (2006) Fuzzy concepts applied to food product quality control:
pH-sensitive (halochromic) smart packaging films based on natu- a review. Fuzzy Sets Syst 157(9):1145–1154. https://​doi.​org/​10.​
ral food colorants for the monitoring of food quality and safety. 1016/j.​fss.​2005.​12.​013
Trends Food Sci Technol 105(January):93–144. https://​doi.​org/​ 39. Doganis P, Alexandridis A, Patrinos P, Sarimveis H (2006) Time
10.​1016/j.​tifs.​2020.​08.​014 series sales forecasting for short shelf-life food products based
22. Halonen N, Pálvölgyi PS, Bassani A, Fiorentini C, Nair R, Spigno on artificial neural networks and evolutionary computing. J Food
G, Kordas K (2020) Bio-based smart materials for food packag- Eng 75(2):196–204. https://​doi.​org/​10.​1016/j.​jfood​eng.​2005.​03.​
ing and sensors – a review. Frontiers in Materials 7(April):1–14. 056
https://​doi.​org/​10.​3389/​fmats.​2020.​00082 40. Szturo K, Szczypinski PM (2017) Ontology based expert system
23. Sun Q, Zhang M, Mujumdar AS (2019) Recent developments of arti- for barley grain classification. Signal Processing - Algorithms,
ficial intelligence in drying of fresh food: a review. Crit Rev Food Architectures, Arrangements, and Applications Conference Pro-
Sci Nutr 59(14):2258–2275. https://​doi.​org/​10.​1080/​10408​398.​2018. ceedings, SPA, 2017-Septe:360–364. https://​doi.​org/​10.​23919/​
​14469​00 SPA.​2017.​81668​93
24. Bhagya Raj GVS, Dash KK (2020) Comprehensive study on appli- 41. Leo Kumar SP (2019) Knowledge-based expert system in manu-
cations of artificial neural network in food process modeling Crit facturing planning: state-of-the-art review. Int J Prod Res 57(15–
Rev Food Sci Nutr 1 28 https://​doi.​org/​10.​1080/​10408​398.​2020.​ 16):4766–4790. https://d​ oi.o​ rg/1​ 0.1​ 080/0​ 02075​ 43.2​ 018.1​ 42437​ 2
18583​98 42. Sipos A (2020) A knowledge-based system as a sustainable soft-
25. Poyatos-Racionero E, Ros-Lis JV, Vivancos JL, Martínez-Máñez ware application for the supervision and intelligent control of
R (2018) Recent advances on intelligent packaging as tools to an alcoholic fermentation process. Sustainability 12(23):10205.
reduce food waste. J Clean Prod 172:3398–3409. https://​doi.​org/​ https://​doi.​org/​10.​3390/​su122​310205
10.​1016/j.​jclep​ro.​2017.​11.​075 43. Ardiansah I, Efatmi F, Mardawati E, Putri SH, Padjadjaran U,
26. Mustafa F, Andreescu S (2018) Chemical and biological sensors Info A, Testing F, Product F, Chaining F, Industries M (2020)
for food-quality monitoring and smart packaging. Foods 7(10). Feasibility testing of a household industry food production cer-
27. Chen S, Brahma S, Mackay J, Cao C, Aliakbarian B (2020) The tificate using an expert system with forward chaining method. J
role of smart packaging system in food supply chain. J Food Sci Inform Frequency 5(2):137–144. https://​doi.​org/​10.​15575/​join.​
85(3):517–525. https://​doi.​org/​10.​1111/​1750-​3841.​15046 v5i2.​579
28. Ahmed I, Lin H, Zou L, Li Z, Brody AL, Qazi IM, Lv L, Pavase TR, 44. Filter M, Appel B, Buschulte A (2015) Expert systems for food
Khan MU, Khan S, Sun L (2018) An overview of smart packaging safety. Curr Opin Food Sci 6:61–65. https://​doi.​org/​10.​1016/j.​
technologies for monitoring safety and quality of meat and meat cofs.​2016.​01.​004
products. Packag Technol Sci 31(7):449–471. https://​doi.​org/​10.​ 45. Skjerdal T, Tessema GT, Fagereng T, Moen LH, Lyshaug L, Gef-
1002/​pts.​2380 ferth A, Spajic M, Estanga EG, De Cesare A, Vitali S, Pasquali
29. Ghoshal G (2018) Recent trends in active, smart, and intelligent F, Bovo F, Manfreda G, Mancusi R, Trevisiani M, Koidis A,
packaging for food products. Elsevier Inc., In Food Packaging Delgado-Pando G, Stratakos AC, Boeri M, Halbert C (2018) The
and Preservation. https://​doi.​org/​10.​1016/​b978-0-​12-​811516-​9.​ STARTEC decision support tool for better tradeoffs between food
00010-5 safety, quality, nutrition, and costs in production of advanced
30. Alam AU, Rathi P, Beshai H, Sarabha GK, Jamal Deen M ready-to-eat foods. Biomed Res Int 2018:1–13
(2021) Fruit quality monitoring with smart packaging. Sen- 46. Suciu I, Ndiaye A, Baudrit C, Fernandez C, Kondjoyan A, Mirade
sors 21(4):1–30. https://​doi.​org/​10.​3390/​s2104​1509 PS, Sicard J, Tournayre P, Bohuon P, Buche P, Courtois F, Guil-
31. Rahman MS, Rashid MM, Hussain MA (2012) Thermal conduc- lard V, Athes V, Flick D, Plana-Fattori A, Trelea C, Trystram G,
tivity prediction of foods by Neural Network and Fuzzy (ANFIS) Delaplace G, Curet S, Della Valle G (2021) A digital learning
modeling techniques. Food Bioprod Process 90(2):333–340. tool based on models and simulators for food engineering (MES-
https://​doi.​org/​10.​1016/j.​fbp.​2011.​07.​001 TRAL). J Food Eng 293(May 2020). https://​doi.​org/​10.​1016/j.​
32. Rahman NA, Hussain MA, Jahim MJ (2012) Production of fruc- jfood​eng.​2020.​110375
tose using recycle fixed-bed reactor and batch bioreactor. J Food 47. Mahdi MS, Ibrahim MF, Mahdi SM, Singam P, Huddin AB
Agric Environ 10(2):268–273 (2019) Fuzzy logic system for diagnosing coronary heart disease.
33. Mozafari MR, Khosravi-Darani K, Borazan GG, Cui J, Pardakhty Int J Eng Technol 8(1.7):119–125.
A, Yurdugul S (2008) Encapsulation of food ingredients using 48. Zadeh LA (2015) Fuzzy logic - a personal perspective. Fuzzy
nanoliposome technology. Int J Food Prop 11(4):833–844. Sets Syst 281:4–20. https://​doi.​org/​10.​1016/j.​fss.​2015.​05.​009
https://​doi.​org/​10.​1080/​10942​91070​16481​15 49. Hannan MA, Ghani ZA, Hoque MM, Ker PJ, Hussain A,
34. Jayasooriya SD, Bhandari BR, Torley P, D’Arcy BR (2004) Effect Mohamed A (2019) Fuzzy logic inverter controller in pho-
of high power ultrasound waves on properties of meat: a review. Int tovoltaic applications: issues and recommendations. IEEE
J Food Prop 7(2):301–319. https://​doi.​org/​10.​1081/​JFP-​12003​0039 Access 7:24934–24955. https://​d oi.​o rg/​1 0.​1 109/​ACCESS.​
35. Saha D, Bhattacharya S (2010) Hydrocolloids as thickening 2019.​28996​10
and gelling agents in food: a critical review. J Food Sci Technol 50. Mutlag AH, Mohamed A, Shareef H (2016) A nature-inspired
47(6):587–597. https://​doi.​org/​10.​1007/​s13197-​010-​0162-6 optimization-based optimum fuzzy logic photovoltaic inverter
36. Belluco S, Losasso C, Maggioletti M, Alonzi CC, Paoletti MG, controller utilizing an eZdsp F28335 board. Energies 9(3).
Ricci A (2013) Edible insects in a food safety and nutritional https://​doi.​org/​10.​3390/​en903​0120
perspective: a critical review. Comprehensive Reviews in Food 51. Rajesh N, Yuh LC, Hashim H, Abd Rahman N, Mohd Ali J
Science and Food Safety 12(3):296–313. https://d​ oi.o​ rg/1​ 0.1​ 111/​ (2021) Food and bioproducts processing fuzzy Mamdani based
1541-​4337.​12014 user-friendly interface for. Food Bioprod Process 126:282–292.
https://​doi.​org/​10.​1016/j.​fbp.​2021.​01.​012
168 Food Engineering Reviews (2022) 14:134–175

52. Alsaqour R, Abdelhaq M, Saeed R, Uddin M, Alsukour O, Mathematics, Engineering and Industrial Applications 2018
Al-Hubaishi M, Alahdal T (2015) Dynamic packet beaconing https://​doi.​org/​10.​1063/1.​50542​21
for GPSR mobile ad hoc position-based routing protocol using 67. Rashmi W, Osama M, Khalid M, Rasheed A, Bhaumik S, Wong
fuzzy logic. J Netw Comput Appl 47:32–46. https://​doi.​org/​10.​ WY, Datta S, Tcsm G (2019) Tribological performance of
1016/j.​jnca.​2014.​08.​008 nanographite-based metalworking fluid and parametric investi-
53. Naf’an E, Universiti KM, Mohamad Ali N, Universiti PIP gation using artificial neural network. Int J Adv Manuf Technol
(2018) Modelling of robot bunker based on fuzzy logic. Digital 104(1–4):359–374. https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00170-0​ 19-0​ 3701-6
Transformation Landscape in the Fourth Industrial Revolution 68. Trafialek J, Laskowski W, Kolanowski W (2015) The use of
(4IR) Era 177–190. Kohonen’s artificial neural networks for analyzing the results of
54. Ali JA, Hannan MA, Mohamed A, Abdolrasol MGM (2016) HACCP system declarative survey. Food Control 51:263–269.
Fuzzy logic speed controller optimization approach for induc- https://​doi.​org/​10.​1016/j.​foodc​ont.​2014.​11.​032
tion motor drive using backtracking search algorithm. Meas.: 69. Said M, Ba-Abbad M, Rozaimah Sheik Abdullah S, Wahab
J  Int Meas Confed 78, 49–62.  https://​ d oi.​ o rg/​ 1 0.​ 1 016/j.​ Mohammad A (2018) Artificial neural network (ANN) for
measu​rement.​2015.​09.​038 optimization of palm oil mill effluent (POME) treatment using
55. Zareiforoush H, Minaei S, Alizadeh MR, Banakar A, Samani reverse osmosis membrane. J Phys Conf Ser 1095(1). https://d​ oi.​
BH (2016) Design, development and performance evaluation org/​10.​1088/​1742-​6596/​1095/1/​012021
of an automatic control system for rice whitening machine 70. da Silva CET, Filardi VL, Pepe IM, Chaves MA, Santos CMS
based on computer vision and fuzzy logic. Comput Electron (2015) Classification of food vegetable oils by fluorimetry and
Agric 124:14–22. https://​doi.​org/​10.​1016/j.​compag.​2016.​01.​ artificial neural networks. Food Control 47:86–91. https://​doi.​
024 org/​10.​1016/j.​foodc​ont.​2014.​06.​030
56. Al-Mahasneh M, Aljarrah M, Rababah T, Alu’datt, M. (2016) 71. Silva SF, Anjos CAR, Cavalcanti RN, Celeghini RMDS (2015)
Application of Hybrid Neural Fuzzy System (ANFIS) in food Evaluation of extra virgin olive oil stability by artificial neural net-
processing and technology. Food Eng Rev 8(3):351–366. work. Food Chem 179:35–43. https://​doi.​org/​10.​1016/j.​foodc​hem.​
https://​doi.​org/​10.​1007/​s12393-​016-​9141-7 2015.​01.​100
57. Baliuta S, Kopylova L, Kuievda I, Kuevda V, Kovalchuk O 72. Shanmuganathan S (2016) Artificial neural network model-
(2020) Fuzzy logic energy management system of food manu- ling: an introduction. Stud Comput Intell (Issue July, pp. 1–14).
facturing processes. Processes and Equipment 9(1):221–239. Springer International Publishing. https://​d oi.​o rg/​1 0.​1 007/​
https://​doi.​org/​10.​24263/​2304- 974X-2020–9–1–19 978-3-​319-​28495-8
58. Cebi N, Sagdic O, Basahel AM, Balubaid MA, Taylan O, 73. Butler KT, Davies DW, Cartwright H, Isayev O, Walsh A (2018)
Yaman M, Yilmaz MT (2019) Modeling and optimization of Machine learning for molecular and materials science. Nature
ultrasound-assisted cinnamon extraction process using fuzzy 559(7715):547–555. https://​doi.​org/​10.​1038/​s41586-​018-0​ 337-2
and response surface models. J Food Process Eng 42(2):1–15. 74. Sharp M, Ak R, Hedberg T (2018) A survey of the advancing use
https://​doi.​org/​10.​1111/​jfpe.​12978 and development of machine learning in smart manufacturing.
59. Kumar K (2016) Intrusion detection using soft computing tech- J Manuf Syst 48:170–179. https://​doi.​org/​10.​1016/j.​jmsy.​2018.​
niques. Int J Comput Commun 6(3):153–169. www.​ijcscn.​com 02.​004
60. Al-Waeli AHA, Sopian K, Yousif JH, Kazem HA, Boland J, 75. Erickson BJ, Korfiatis P, Akkus Z, Kline TL (2017) Machine
Chaichan MT (2019) Artificial neural network modeling and learning for medical imaging. Radiographics 37(2):505–515.
analysis of photovoltaic/thermal system based on the experi- https://​doi.​org/​10.​1148/​rg.​20171​60130
mental study. Energy Convers Manag 186(November 2018), 76. Mullainathan S, Spiess J (2017) Machine learning: an applied
368–379. https://​doi.​org/​10.​1016/j.​encon​man.​2019.​02.​066 econometric approach. Journal of Economic Perspectives
61. Gandhi N, Petkar O, Armstrong LJ (2016) Rice crop yield 31(2):87–106. https://​doi.​org/​10.​1257/​jep.​31.2.​87
prediction using artificial neural networks. Proceedings - 2016 77. Deo RC (2015) Machine learning in medicine. Circulation
IEEE International Conference on Technological Innovations 132(20):1920–1930. https://​doi.​org/​10.​1161/​CIRCU​LATIO​NAHA.​
in ICT for Agriculture and Rural Development, TIAR 2016, 115.​001593
105–110. https://​doi.​org/​10.​1109/​TIAR.​2016.​78012​22 78. Carleo G, Cirac I, Cranmer K, Daudet L, Schuld M, Tishby N,
62. Gonzalez-Fernandez I, Iglesias-Otero MA, Esteki M, Moldes Vogt-Maranto L, Zdeborová L (2019) Machine learning and
OA, Mejuto JC, Simal-Gandara J (2019) A critical review on the physical sciences. Rev Mod Phys 91(4):45002. https://​doi.​
the use of artificial neural networks in olive oil production, org/​10.​1103/​RevMo​dPhys.​91.​045002
characterization and authentication. Crit Rev Food Sci Nutr 79. Jordan MI, Mitchell TM (2015) Machine learning: trends, per-
59(12):1913–1926. https://​doi.​org/​10.​1080/​10408​398.​2018.​ spectives, and prospects. Science 349(6245):255–260. https://​
14336​28 doi.​org/​10.​1126/​scien​ce.​aaa84​15
63. Maladkar K (2018) 6 types of artificial neural networks cur- 80. Rajkomar A, Dean J, Kohane I (2019) Machine learning in medi-
rently being used in machine learning. Retrieved from: https://​ cine. N Engl J Med 380(14):1347–1358. https://​doi.​org/​10.​1056/​
analy​ticsi​ndiam​ag.​com/6-​types-​of-​artif​i cial-​neural-​netwo​rks-​ NEJMr​a1814​259
curre​ntly-​being-​used-​in-​todays-​techn​ologi​es/ 81. Estelles-Lopez L, Ropodi A, Pavlidis D, Fotopoulou J, Gkousari
64. Abdul Aziz FAB, Rahman N, Mohd Ali J (2019) Tropospheric C, Peyrodie A, Panagou E, Nychas GJ, Mohareb F (2017) An
ozone formation estimation in Urban City, Bangi, Using Artifi- automated ranking platform for machine learning regression
cial Neural Network (ANN). Comput Intell Neurosci 2019:1–10. models for meat spoilage prediction using multi-spectral imag-
https://​doi.​org/​10.​1155/​2019/​62529​83 ing and metabolic profiling. Food Res Int 99:206–215. https://​
65. Khamis N, Mat Yazid MR, Hamim A, Rosyidi SAP, Nur NI, doi.​org/​10.​1016/j.​foodr​es.​2017.​05.​013
Borhan MN (2018) Predicting the rheological properties of 82. Lu NV, Vuong TN, Dinh DT (2020) Combining correlation-
bitumen-filler mastic using artificial neural network methods. J based feature and machine learning for sensory evaluation of
Teknol 80(1):71–78. https://​doi.​org/​10.​11113/​jt.​v80.​11097 saigon beer. International Journal of Knowledge and Systems
66. Ismail M, Jubley NZ, Ali ZM (2018) Forecasting Malaysian for- Science 11(2):71–85. https://d​ oi.o​ rg/1​ 0.4​ 018/I​ JKSS.2​ 02004​ 0104
eign exchange rate using artificial neural network and ARIMA 83. Li B, Lin Y, Yu W, Wilson DI, Young BR (2020) Application of
time series Proceeding of the International Conference on mechanistic modelling and machine learning for cream cheese
Food Engineering Reviews (2022) 14:134–175 169

fermentation pH prediction. J Chem Technol Biotechnol. https://​ 99. Baietto M, & Wilson AD (2015) Electronic-nose applications for
doi.​org/​10.​1002/​jctb.​6517 fruit identification, ripeness and quality grading. In Sensors (Swit-
84. Kim DH, Zohdi TI, Singh RP (2020) Modeling, simulation and zerland) 15(1):899–931. https://​doi.​org/​10.​3390/​s1501​00899
machine learning for rapid process control of multiphase flowing 100. Gliszczyńska-Świgło A, Chmielewski J (2017) Electronic nose
foods. Comput Methods Appl Mech Eng 371:113286. https://d​ oi.​ as a tool for monitoring the authenticity of food. A Review
org/​10.​1016/j.​cma.​2020.​113286 Food Analytical Methods 10(6):1800–1816. https://​doi.​org/1​ 0.​
85. Alaiz-Rodriguez R, Parnell AC (2020) A machine learning 1007/​s12161-​016-​0739-4
approach for lamb meat quality assessment using FTIR spectra. 101. Mohamed RR, Taacob R, Mohamed MA, Tengku Dir TA, Rahim
IEEE Access 8:52385–52394. https://​doi.​org/​10.​1109/​ACCESS.​ FA, Mamat AR (2015) Data mining techniques in food safety. Int
2020.​29746​23 J Adv Trends Comput Sci Eng 9(1.1):379–384.
86. Tsoumakas G (2019) A survey of machine learning techniques 102. Ha D, Sun Q, Su K, Wan H, Li H, Xu N, Sun F, Zhuang L,
for food sales prediction. Artif Intell Rev 52(1):441–447. https://​ Hu N, Wang P (2015) Recent achievements in electronic tongue
doi.​org/​10.​1007/​s10462-​018-​9637-z and bioelectronic tongue as taste sensors. Sensor Actuat B-Chem
87. Garre A, Ruiz MC, Hontoria E (2020) Application of machine 207(PB):1136–1146. https://​doi.​org/​10.​1016/j.​snb.​2014.​09.​077
learning to support production planning of a food industry in 103. Podrazka M, Báczyńska E, Kundys M, Jeleń PS, Nery EW (2017)
the context of waste generation under uncertainty. Operations Electronic tongue-a tool for all tastes? Biosensors 8(1):1–24.
Research Perspectives 7(January):100147. https://​doi.​org/​10.​ https://​doi.​org/​10.​3390/​bios8​010003
1016/j.​orp.​2020.​100147 104. Lan Y, Wu J, Wang X, Sun X, Hackman RM, Li Z, Feng X (2017)
88. Melin P, Miramontes I, Prado-Arechiga G (2018) A hybrid model Evaluation of antioxidant capacity and flavor profile change of
based on modular neural networks and fuzzy systems for classifi- pomegranate wine during fermentation and aging process. Food
cation of blood pressure and hypertension risk diagnosis. Expert Chem 232:777–787. https://d​ oi.o​ rg/1​ 0.1​ 016/j.f​ oodch​ em.​2017.​04.​
Syst Appl 107:146–164. https://​doi.​org/​10.​1016/j.​eswa.​2018.​04.​ 030
023 105. Phat C, Moon B, Lee C (2016) Evaluation of umami taste in
89. Mamat RC, Kasa A, Razali SFM, Samad AM, Ramli A, Yazid mushroom extracts by chemical analysis, sensory evaluation, and
MRM (2019) Application of artificial intelligence in predicting an electronic tongue system. Food Chem 192:1068–1077. https://​
ground settlement on earth slope. AIP Conf Proc 2138(August). doi.​org/​10.​1016/j.​foodc​hem.​2015.​07.​113
https://​doi.​org/​10.​1063/1.​51210​94 106. Jiang H, Zhang M, Bhandari B, Adhikari B (2018) Application
90. Bouhoune K, Yazid K, Boucherit MS, Chériti A (2017) Hybrid of electronic tongue for fresh foods quality evaluation: a review.
control of the three phase induction machine using artificial neural Food Rev Intl 34(8):746–769. https://d​ oi.o​ rg/1​ 0.1​ 080/8​ 75591​ 29.​
networks and fuzzy logic. Applied Soft Computing Journal 55:289– 2018.​14241​84
301. https://​doi.​org/​10.​1016/j.​asoc.​2017.​01.​048 107. Cárdenas-Pérez S, Chanona-Pérez J, Méndez-Méndez JV,
91. Sharma LK, Vishal V, Singh TN (2017) Developing novel mod- Calderón-Domínguez G, López-Santiago R, Perea-Flores MJ,
els using neural networks and fuzzy systems for the prediction Arzate-Vázquez I (2017) Evaluation of the ripening stages of
of strength of rocks from key geomechanical properties. Meas.: apple (Golden Delicious) by means of computer vision system.
J Int Meas Confed  102, 158–169. https://​doi.​org/​10.​1016/j.​ Biosys Eng 159:46–58. https://​doi.​org/​10.​1016/j.​biosy​stems​eng.​
measu​rement.​2017.​01.​043 2017.​04.​009
92. Viharos ZJ, Kis KB (2015) Survey on Neuro-Fuzzy systems and 108. De Oliveira EM, Leme DS, Barbosa BHG, Rodarte MP, Alva-
their applications in technical diagnostics and measurement. renga Pereira RGF (2016) A computer vision system for cof-
Meas.: J Int Meas Confed 67, 126–136. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​ fee beans classification based on computational intelligence
measu​rement.​2015.​02.​001 techniques. J Food Eng 171:22–27. https://​doi.​org/​10.​1016/j.
93. Ali JA, Hannan MA, Mohamed A, Humaidi AJ (2015) Adap- ​jfood​eng.​2015.​10.​009
tive neuro fuzzy inference system-based space vector PWM 109. Sun X, Young J, Liu JH, Bachmeier L, Somers RM, Chen KJ,
inverter for three-phase induction motor drive. 5th Int Conf Newman D (2016) Prediction of pork color attributes using
Electr Electron Eng ICEEE 238–243. computer vision system. Meat Sci 113:62–64. https://​doi.​org/​10.​
94. Yan J, Guo X, Duan S, Jia P, Wang L, Peng C, Zhang S (2015) 1016/j.​meats​ci.​2015.​11.​009
Electronic nose feature extraction methods: a review. Sensors 110. Leme DS, Da Silva SA, Barbosa BHG, Borém FM, Pereira
(Switzerland) 15(11):27804–27831. https://​doi.​org/​10.​3390/​ RGFA (2019) Recognition of coffee roasting degree using a
s1511​27804 computer vision system. Comput Electron Agric 156(October
95. Deshmukh S, Bandyopadhyay R, Bhattacharyya N, Pandey RA, 2018):312–317. https://​doi.​org/​10.​1016/j.​compag.​2018.​11.​029
Jana A (2015) Application of electronic nose for industrial odors 111. Cavallo D, Pietro Cefola M, Pace B, Logrieco AF, Attolico G
and gaseous emissions measurement and monitoring - an over- (2019) Non-destructive and contactless quality evaluation of
view. Talanta 144:329–340. https://​doi.​org/​10.​1016/j.​t alan​t a. table grapes by a computer vision system. Comput Electron
​2015.​06.​050 Agric 156(June 2018):558–564. https://​doi.​org/​10.​1016/j.​com-
96. Szulczyński B, Gȩbicki J, Namieśnik J (2018) Application pag.​2018.​12.​019
of fuzzy logic to determine the odour intensity of model gas 112. Chmiel M, Słowiński M (2016) The use of computer vision sys-
mixtures using electronic nose. E3S Web Conf 28(2):15–21. tem to detect pork defects. LWT Food Sci Technol 73:473–480.
https://​doi.​org/​10.​1051/​e3sco​nf/​20182​801036 https://​doi.​org/​10.​1016/j.​lwt.​2016.​06.​054
97. Wojnowski W, Majchrzak T, Dymerski T, Gębicki J, Namieśnik 113. Columbus L (2020) 10 ways AI is improving manufacturing in
J (2017) Portable electronic nose based on electrochemical 2020. Retrieved from 10 Ways AI Is Improving Manufacturing
sensors for food quality assessment. Sensors (Switzerland) In 2020 (forbes.com).
17(12):1–14. https://​doi.​org/​10.​3390/​s1712​2715 114. Qu JH, Liu D, Cheng JH, Sun DW, Ma J, Pu H, Zeng XA (2015)
98. Sanaeifar A, ZakiDizaji H, Jafari A, de la Guardia M (2017) Early Applications of near-infrared spectroscopy in food safety evaluation
detection of contamination and defect in foodstuffs by electronic and control: a review of recent research advances. Crit Rev Food
nose: a review. TrAC - Trends in Analytical Chemistry 97:257–271. Sci Nutr 55(13):1939–1954. https://​doi.​org/​10.​1080/​10408​398.
https://​doi.​org/​10.​1016/j.​trac.​2017.​09.​014 ​2013.​871693
170 Food Engineering Reviews (2022) 14:134–175

115. Porep JU, Kammerer DR, Carle R (2015) On-line application Innov Manag Rev 7(11):12–20. https://​doi.​org/​10.​22215/​timre​view/​
of near infrared (NIR) spectroscopy in food production. Trends 1323
Food Sci Technol 46(2):211–230. https://​doi.​org/​10.​1016/j.​tifs.​ 131. Bai C, Dallasega P, Orzes G, & Sarkis J (2020) Industry 4.0 tech-
2015.​10.​002 nologies assessment: a sustainability perspective. Int J Prod Econ
116. Fu X, Ying Y (2016) Food safety evaluation based on near infra- 229, 107776. https://​doi.​org/​10.​1016/j.​ijpe.​2020.​107776
red spectroscopy and imaging: a review. Crit Rev Food Sci Nutr 132. Pan Y (2016) Heading toward artificial intelligence 2.0. Engi-
56(11):1913–1924. https://​doi.​org/​10.​1080/​10408​398.​2013.​ neering 2(4):409–413. https://​doi.​org/​10.​1016/J.​ENG.​2016.​04.​
807418 018
117. Cozzolino D (2016) Near infrared spectroscopy and food authentic- 133. Budiyanto G, Ipnuwati S, Al Gifari SA, Huda M, Jalal B, Abdul
ity Advances in Food Traceability Techniques and Technologies Latif A, Lia Hananto A (2018) Web based expert system for
119–136 https://​doi.​org/​10.​1016/​B978-0-​08-​100310-​7.​00007-7 diagnosing disease pest on banana plant.  Int J  Eng Technol
118. Grassi S, Alamprese C (2018) Advances in NIR spectroscopy (UAE) 7(4):4715–4721. https://​doi.​org/​10.​14419/​ijet
applied to process analytical technology in food industries. Curr 134. Hernández-Vera B, Aguilar Lasserre AA, Gastón Cedillo-
Opin Food Sci 22:17–21. https://​doi.​org/​10.​1016/j.c​ ofs.​2017.1​ 2.​ Campos M, Herrera-Franco LE,  Ochoa-Robles J (2017a) Expert
008 system based on fuzzy logic to define the production process in
119. Cortés V, Blasco J, Aleixos N, Cubero S, Talens P (2019) Monitoring the coffee industry. J Food Process Eng 40(2). https://​doi.​org/​
strategies for quality control of agricultural products using visible 10.​1111/​jfpe.​12389
and near-infrared spectroscopy: A review. Trends Food Sci Technol 135. Livio J, Hodhod R (2018) AI cupper: a fuzzy expert system for
85(October 2018):138–148. https://​doi.​org/​10.​1016/j.​tifs.​2019.​01.​ sensorial evaluation of coffee bean attributes to derive quality
015 scoring. IEEE Trans Fuzzy Syst 26(6):3418–3427. https://​doi.​
120. Núñez-Sánchez N, Martínez-Marín AL, Polvillo O, Fernández- org/​10.​1109/​TFUZZ.​2018.​28326​11
Cabanás VM, Carrizosa J, Urrutia B, Serradilla JM (2016) Near 136. Sumaryanti L, Istanto T, Pare S (2020) Rule based method in
Infrared Spectroscopy (NIRS) for the determination of the milk expert system for detection pests and diseases of corn. J Phys
fat fatty acid profile of goats. Food Chem 190:244–252. https://​ Conf Ser 1569(2). https://​doi.​org/​10.​1088/​1742-​6596/​1569/2/​
doi.​org/​10.​1016/j.​foodc​hem.​2015.​05.​083 022023
121. Kartakoullis A, Comaposada J, Cruz-Carrión A, Serra X, Gou 137. Zakaria MZ, Nordin N, Malik AMA, Elias SJ, Shahuddin AZ
P (2019) Feasibility study of smartphone-based Near Infrared (2019) Fuzzy expert systems (FES) for halal food additive.
Spectroscopy (NIRS) for salted minced meat composition diag- Indones J Electr Eng Comput Sci 13(3):1073–1078. https://​doi.​
nostics at different temperatures. Food Chem 278:314–321. org/​10.​11591/​ijeecs.​v13.​i3.​pp1073-​1078
https://​doi.​org/​10.​1016/j.​foodc​hem.​2018.​11.​054 138. Blagoveshchenskiy IG, Blagoveshchenskiy VG, Besfamilnaya
122. De Marchi M, Manuelian CL, Ton S, Manfrin D, Meneghesso EM, Sumerin VA (2020) Development of databases of intelligent
M, Cassandro M, Penasa M (2017) Prediction of sodium content expert systems for automatic control of product quality indica-
in commercial processed meat products using near infrared spec- tors. J Phys Conf Ser 1705(1). https://​doi.​org/​10.​1088/​1742-​6596/​
troscopy. Meat Sci 125:61–65. https://​doi.​org/1​ 0.​1016/j.m ​ eatsc​ i.​ 1705/1/​012019
2016.​11.​014 139. Bortolini M, Faccio M, Ferrari E, Gamberi M, Pilati F (2016)
123. Wold JP, Veiseth-Kent E, Høst V, Løvland A (2017) Rapid on- Fresh food sustainable distribution: cost, delivery time and carbon
line detection and grading of wooden breast myopathy in chicken footprint three-objective optimization. J Food Eng 174:56–67.
fillets by near-infrared spectroscopy. PLoS ONE 12(3):1–16. https://​doi.​org/​10.​1016/j.​jfood​eng.​2015.​11.​014
https://​doi.​org/​10.​1371/​journ​al.​pone.​01733​84 140. Vásquez RP, Aguilar-Lasserre AA, López-Segura MV, Rivero
124. Olarewaju OO, Bertling I, Magwaza LS (2016) Non-destructive LC, Rodríguez-Duran AA, Rojas-Luna MA (2019) Expert sys-
evaluation of avocado fruit maturity using near infrared spec- tem based on a fuzzy logic model for the analysis of the sus-
troscopy and PLS regression models. Sci Hortic 199:229–236. tainable livestock production dynamic system. Comput Electron
https://​doi.​org/​10.​1016/j.​scien​ta.​2015.​12.​047 Agric 161(January):104–120. https://​doi.​org/​10.​1016/j.​compag.​
125. Adedipe OE, Johanningsmeier SD, Truong VD, Yencho GC 2018.​05.​015
(2016) Development and validation of a near-infrared spectros- 141. Nicolotti L, Mall V, Schieberle P (2019) Characterization of
copy method for the prediction of acrylamide content in French- key aroma compounds in a commercial rum and an Australian
fried potato. J Agric Food Chem 64(8):1850–1860. https://​doi.​ red wine by means of a new Sensomics-Based Expert System
org/​10.​1021/​acs.​jafc.​5b047​33 (SEBES) - an approach to use artificial intelligence in determin-
126. Pullanagari RR, Yule IJ, Agnew M (2015) On-line prediction of ing food odor codes. J Agric Food Chem 67(14):4011–4022.
lamb fatty acid composition by visible near infrared spectros- https://​doi.​org/​10.​1021/​acs.​jafc.​9b007​08
copy. Meat Sci 100:156–163. https://​doi.​org/​10.​1016/j.​meats​ci.​ 142. Kharisma Adi K, Isnanto RR (2020) Rice crop management
2014.​10.​008 expert system with forwarding chaining method and certainty
127. Liang N, Sun S, Zhang C, He Y, Qiu Z (2020) Advances in factor. J Phys Conf Ser 1524(1). https://​doi.​org/​10.​1088/​1742-​
infrared spectroscopy combined with artificial neural network 6596/​1524/1/​012037
for the authentication and traceability of food Crit Rev Food Sci 143. Rajendra L, Azani H, Much I, Subroto I, Marwanto A (2017)
Nutr 1–22 https://​doi.​org/​10.​1080/​10408​398.​2020.​18620​45 Expert system on soybean disease using knowledge represen-
128. Tan J, Xu J (2020) Applications of electronic nose (e-nose) and tation method. Telemat Inform 5(1):36–46. https://​doi.​org/​10.​
electronic tongue (e-tongue) in food quality-related properties 12928/​jti.​v5i1
determination: A review. Artificial Intelligence in Agriculture 144. Lamastra L, Balderacchi M, Di Guardo A, Monchiero M,
4:104–115. https://​doi.​org/​10.​1016/j.​aiia.​2020.​06.​003 Trevisan M (2016) A novel fuzzy expert system to assess the
129. Stăncioiu A (2017) The Fourth Industrial Revolution “Industry sustainability of the viticulture at the wine-estate scale. Sci Total
4.0.” Fiabilitate Şi Durabilitate 1, 74–78. http://​www.​utgjiu.​ro/​ Environ 572:724–733. https://​doi.​org/​10.​1016/j.​scito​tenv.​2016.​
rev_​mec/​mecan​ica/​pdf/​2017-​01/​11_​Alin STĂNCIOIU - THE 07.​043
FOURTH INDUSTRIAL REVOLUTION INDUSTRY 4.0”.pdf 145. Chowdhury T, Das M (2015) Sensory evaluation of aromatic
130. Morrar R, Arman H, Mousa S (2017) The fourth industrial revo- foods packed in developed starch based films using fuzzy logic.
lution (Industry 4.0): a social innovation perspective. Technol
Food Engineering Reviews (2022) 14:134–175 171

International Journal of Food Studies 4(1):29–48. https://d​ oi.o​ rg/​ 221(November):936–943. https://​doi.​org/​10.​1016/j.​foodc​hem.​
10.​7455/​ijfs.​v4i1.​228 2016.​11.​071
146. Fatma S, Sharma N, Singh SP, Jha A, Kumar A (2016) Fuzzy 162. Basak S (2018) The use of fuzzy logic to determine the con-
analysis of sensory data for ranking of beetroot candy. Int J Food centration of betel leaf essential oil and its potency as a juice
Eng 2(1):26–30. https://​doi.​org/​10.​18178/​ijfe.2.​1.​26-​30 preservative. Food Chem 240(August 2017): 1113–1120. https://​
147. Chung CC, Chen HH, Ting CH (2016) Fuzzy logic for accu- doi.​org/​10.​1016/j.​foodc​hem.​2017.​08.​047
rate control of heating temperature and duration in canned food 163. Jahedi Rad S, Kaveh M, Sharabiani VR, Taghinezhad E (2018)
sterilisation. Engineering in Agriculture, Environment and Food Fuzzy logic, artificial neural network and mathematical model for
9(2):187–194. https://​doi.​org/​10.​1016/j.​eaef.​2015.​11.​003 prediction of white mulberry drying kinetics. Heat Mass Transf
148. Hernández-Vera B, Aguilar Lasserre AA, Gastón Cedillo- 54(11):3361–3374. https://​doi.​org/​10.​1007/​s00231-​018-​2377-4
Campos M, Herrera-Franco LE, Ochoa-Robles J (2017) Expert 164. Benković M, Tušek AJ, Belščak-Cvitanović A, Lenart A, Domian
system based on fuzzy logic to define the production process in E, Komes D, Bauman I (2015) Artificial neural network mod-
the coffee industry. J Food Process Eng 40(2):1–10. https://​doi.​ elling of changes in physical and chemical properties of cocoa
org/​10.​1111/​jfpe.​12389 powder mixtures during agglomeration. LWT Food Sci Technol
149. Harsawardana Samodro B, Mahesworo B, Suparyanto T, Surya 64(1):140–148. https://​doi.​org/​10.​1016/j.​lwt.​2015.​05.​028
Atmaja DB, Pardamean B (2020) Maintaining the quality and 165. Koszela K, Łukomski M, Mueller W, Górna K, Okoń P, Boniecki
aroma of coffee with fuzzy logic coffee roasting machine. IOP P, Zaborowicz M, Wojcieszak D (2017) Classification of dried
Conf Ser Earth Environ Sci 426(1). https://d​ oi.o​ rg/1​ 0.1​ 088/1​ 755-​ vegetables using computer image analysis and artificial neural
1315/​426/1/​012148 networks. Ninth International Conference on Digital Image Pro-
150. Singh V, Kumar S, Singh J, Rai AK (2018) Fuzzy logic sen- cessing (ICDIP 2017), 10420(Icdip), 1042031. https://​doi.​org/​
sory evaluation of cupcakes developed from the Mahua flower 10.​1117/​12.​22817​18
( Madhuca Longifolia ). Journal of Emerging Technologies and 166. Bahmani A, Jafari SM, Shahidi S-A, Dehnad D (2015) Mass
Innovative Research 5(1):411–421 transfer kinetics of eggplant during osmotic dehydration by neu-
151. Yousefi-Darani A, Paquet-Durand O, Hitzmann B (2019) Appli- ral networks. J Food Process Preserv 1:1–13. https://​doi.​org/​10.​
cation of fuzzy logic control for the dough proofing process. Food 1111/​jfpp.​12435
Bioprod Process 115:36–46. https://​doi.​org/​10.​1016/j.​fbp.​2019.​ 167. Liu J, Liu L, Guo W, Fu M, Yang M, Huang S, Zhang F, Liu Y
02.​006 (2019) A new methodology for sensory quality assessment of gar-
152. Farzaneh V, Bakhshabadi H, Gharekhani M, Ganje M, Farzaneh lic based on metabolomics and an artificial neural network. RSC
F, Rashidzadeh S, Carvalho SI (2017) Application of an adap- Adv 9(31):17754–17765. https://​doi.​org/​10.​1039/​c9ra0​1978b
tive neuro_fuzzy inference system (ANFIS) in the modeling of 168. Naik RR, Gandhi NS, Thakur M, Nanda V (2019) Analysis of
rapeseeds’ oil extraction. J Food Process Eng 40(6):1–8. https://​ crystallization phenomenon in Indian honey using molecular
doi.​org/​10.​1111/​jfpe.​12562 dynamics simulations and artificial neural network. Food Chem
153. Farzaneh V, Ghodsvali A, Bakhshabadi H, Ganje M, Dolatabadi Z, 300(1):125182. https://d​ oi.o​ rg/1​ 0.1​ 016/j.f​ oodch​ em.2​ 019.1​ 25182
Carvalho S, I. (2017) Modelling of the selected physical properties 169. Dang NT, Vo MT, Nguyen TD, Dao SVT (2019) Analysis on
of the fava bean with various moisture contents using fuzzy logic mangoes weight estimation problem using neural network. Pro-
design. J Food Process Eng 40(2):1–9. https://​doi.​org/​10.​1111/​jfpe.​ ceedings - 2019 19th International Symposium on Communica-
12366 tions and Information Technologies, ISCIT 2019, i, 559–562.
154. Shahidi B, Sharifi A, Roozbeh Nasiraie L, Niakousari M, Ahmadi M https://​doi.​org/​10.​1109/​ISCIT.​2019.​89051​18
(2020) Phenolic content and antioxidant activity of flixweed (Descu- 170. Omari A, Behroozi-Khazaei N, Sharifian F (2018) Drying kinetic
rainia sophia) seeds extracts: ranking extraction systems based on fuzzy and artificial neural network modeling of mushroom drying pro-
logic method. Sustain Chem Pharm 16(March). cess in microwave-hot air dryer. J Food Process Eng 41(7):1–10.
155. Kaushik N, Gondi AR, Rana R, Srinivasa Rao P (2015) Applica- https://​doi.​org/​10.​1111/​jfpe.​12849
tion of fuzzy logic technique for sensory evaluation of high pres- 171. Ardabili S, Mosavi A, Mahmoudi A, Gundoshmian Tarahom
sure processed mango pulp and litchi juice and its comparison to Mesri Nosratabadi S, Várkonyi-Kóczy AR (2020) Modelling
thermal treatment. In Innov Food Sci Emerg (Vol. 32). Elsevier temperature variation of mushroom growing hall using artificial
B.V. https://​doi.​org/​10.​1016/j.​ifset.​2015.​08.​007 neural networks. In J Sustain Dev (101). Springer, Cham. https://​
156. Amiryousefi MR, Mohebbi M, Golmohammadzadeh S, Koocheki doi.​org/​10.​1007/​978-3-​030-​36841-8_​10
A, Baghbani F (2017) Fuzzy logic application to model caffeine 172. Azadbakht M, Aghili H, Ziaratban A, Torshizi MV (2017)
release from hydrogel colloidosomes. J Food Eng 212:181–189. Application of artificial neural network method to exergy and
https://​doi.​org/​10.​1016/j.​jfood​eng.​2017.​05.​031 energy analyses of fluidized bed dryer for potato cubes. Energy
157. Jafari SM, Ganje M, Dehnad D, Ghanbari V (2016) Mathemati- 120:947–958. https://​doi.​org/​10.​1016/j.​energy.​2016.​12.​006
cal, fuzzy logic and artificial neural network modeling techniques 173. Anastácio A, Silva R, Carvalho IS (2016) Phenolics extrac-
to predict drying kinetics of onion. J Food Process Preserv tion from sweet potato peels: modelling and optimization by
40(2):329–339. https://​doi.​org/​10.​1111/​jfpp.​12610 response surface modelling and artificial neural network. J
158. Sarkar T, Bhattacharjee R, Salauddin M, Giri A, Chakraborty R Food Sci Technol 53(12):4117–4125. https://​doi.​org/​10.​1007/​
(2020) Application of fuzzy logic analysis on pineapple Rasgulla. s13197-​016-​2354-1
Procedia Computer Science 167(2019):779–787. https://​doi.​org/​ 174. Chasiotis VK, Tzempelikos DA, Filios AE, Moustris KP (2020)
10.​1016/j.​procs.​2020.​03.​410 Artificial neural network modelling of moisture content evolution for
159. Blasi A (2018) Scheduling food industry system using fuzzy convective drying of cylindrical quince slices. Comput Electron Agric
logic. J Theor Appl Inf Technol 96(19):6463–6473 172(June):105074. https://​doi.​org/​10.​1016/j.​compag.​2019.​105074
160. Yulianto T, Komariyah S, Ulfaniyah N (2017) Application of fuzzy 175. Chen Y, Cai K, Tu Z, Nie W, Ji T, Hu B, Chen C (2018) Pre-
inference system by Sugeno method on estimating of salt produc- diction of benzo[a]pyrene content of smoked sausage using
tion AIP Conf Proc 1867 https://​doi.​org/​10.​1063/1.​49944​42 back-propagation artificial neural network. J  Sci Food Agr
161. Zare D, Ghazali HM (2017) Assessing the quality of sardine 98(8):3032–3030. https://​doi.​org/​10.​1002/​jsfa.​8801
based on biogenic amines using a fuzzy logic model. Food Chem 176. Li M, Ekramirad N, Rady A, Adedeji A (2018) Application of
acoustic emission and machine learning to detect codling moth
172 Food Engineering Reviews (2022) 14:134–175

infested apples. ASABE 61(3):1157–1164. https://​doi.​org/​10.​ 190. Abbaspour-Gilandeh Y, Jahanbakhshi A, Kaveh M (2020) Pre-
13031/​trans.​12548 1157 diction kinetic, energy and exergy of quince under hot air dryer
177. Sabater C, Olano A, Corzo N, Montilla A (2019) GC–MS using ANNs and ANFIS. Food Sci Nutr 8(1):594–611. https://​
characterisation of novel artichoke (Cynara scolymus) pectic- doi.​org/​10.​1002/​fsn3.​1347
oligosaccharides mixtures by the application of machine learning 191. Soleimanzadeh B, Hemati L, Yolmeh M, Salehi F (2015) GA-
algorithms and competitive fragmentation modelling. Carbohyd ANN and ANFIS models and salmonella enteritidis inactivation
Polym 205:513–523. https://​doi.​org/​10.​1016/j.​carbp​ol.​2018.​10.​ by ultrasound. J Food Saf 35(2):220–226. https://​doi.​org/​10.​
054 1111/​jfs.​12174
178. Claudia Gonzalez FRV, Sigfredo F, Damir T, Kate H, Dunshea 192. Kumar V, Sharma HK (2017) Process optimization for extrac-
(2017) Assessment of beer quality based on foamability and tion of bioactive compounds from taro (Colocasia esculenta),
chemical composition using computer vision algorithms, near using RSM and ANFIS modeling. Journal of Food Measure-
infrared spectroscopy and machine learning algorithms. J Sci ment and Characterization 11(2):704–718. https://​doi.​org/​10.​
Food Agr 1–39. https://​doi.​org/​10.​1002/​jsfa.​8506 1007/​s11694-​016-​9440-y
179. De Sousa Silva M, Cruz LF, Bugatti PH, Saito PTM (2020) 193. Kaveh M, Rasooli Sharabiani V, Amiri Chayjan R, Taghinezhad
Automatic visual quality assessment of biscuits using machine E, Abbaspour-Gilandeh Y, Golpour I (2018) ANFIS and ANNs
learning. In L. Rutkowski, M. Scherer Rafałand Korytkowski, W. model for prediction of moisture diffusivity and specific energy
Pedrycz, R. Tadeusiewicz, & J. M. Zurada (Eds.), J Artif Intell consumption potato, garlic and cantaloupe drying under con-
Soft (pp. 59–70). Springer International Publishing. vective hot air dryer. Information Processing in Agriculture
180. Younis K, Ahmad S, Osama K, Malik MA (2019) Optimiza- 5(3):372–387. https://​doi.​org/​10.​1016/j.​inpa.​2018.​05.​003
tion of de-bittering process of mosambi (Citrus limetta) peel: 194. Arabameri M, Nazari RR, Abdolshahi A, Abdollahzadeh
Artificial neural network, Gaussian process regression and sup- M, Mirzamohammadi S, Shariatifar N, Barba FJ, Mousavi
port vector machine modeling approach. J Food Process Eng Khaneghah A (2019) Oxidative stability of virgin olive oil:
42(6):1–12. https://​doi.​org/​10.​1111/​jfpe.​13185 evaluation and prediction with an adaptive neuro-fuzzy infer-
181. Astray G, Albuquerque BR, Prieto MA, Simal-Gandara J, Fer- ence system (ANFIS). J Sci Food Agric 99(12):5358–5367.
reira ICFR, Barros L (2020) Stability assessment of extracts https://​doi.​org/​10.​1002/​jsfa.​9777
obtained from Arbutus unedo L. fruits in powder and solution 195. Ojediran JO, Okonkwo CE, Adeyi AJ, Adeyi O, Olaniran AF,
systems using machine-learning methodologies. Food Chem George NE, Olayanju AT (2020) Drying characteristics of yam
333(January) 127460. https://​doi.​org/​10.​1016/j.​foodc​hem.​2020.​ slices (Dioscorea rotundata) in a convective hot air dryer: appli-
127460 cation of ANFIS in the prediction of drying kinetics. Heliyon
182. Pise D, Upadhye GD (2018) Grading of harvested mangoes 6(3):e03555. https://​doi.​org/​10.​1016/j.​heliy​on.​2020.​e03555
quality and maturity based on machine learning techniques. 196. Kodogiannis VS, Alshejari A (2016) Neuro-fuzzy based iden-
2018 Int Conf Smart City Emerg Technol ICSCET 2018 1–6. tification of meat spoilage using an electronic nose. 2016 IEEE
https://​doi.​org/​10.​1109/​ICSCET.​2018.​85373​42 8th International Conference on Intelligent Systems, IS 2016 -
183. Gutiérrez P, Godoy SE, Torres S, Oyarzún P, Sanhueza I, Díaz- Proceedings, 96–103. https://​doi.​org/​10.​1109/​IS.​2016.​77374​06
García V, Contreras-Trigo B, Coelho P (2020) Improved anti- 197. Tan J, Balasubramanian B, Sukha D, Ramkissoon S, Umaharan
biotic detection in raw milk using machine learning tools over P (2019) Sensing fermentation degree of cocoa (Theobroma
the absorption spectra of a problem-specific nanobiosensor. cacao L.) beans by machine learning classification models
Sensors (Switzerland) 20(16):1–13. https://​doi.​org/​10.​3390/​ based electronic nose system. J Food Process Eng 42(6):1–8.
s2016​4552 https://​doi.​org/​10.​1111/​jfpe.​13175
184. Xu JL, Sun DW (2017) Identification of freezer burn on fro- 198. Thazin Y, Pobkrut T, Kerdcharoen T (2018) Prediction of acidity
zen salmon surface using hyperspectral imaging and com- levels of fresh roasted coffees using E-nose and artificial neural
puter vision combined with machine learning algorithm network. 2018 10th International Conference on Knowledge and
d’apprentissage automatique. Int J Refrig 74:149–162. https://​ Smart Technology: Cybernetics in the Next Decades, KST 2018,
doi.​org/​10.​1016/j.​ijref​r ig.​2016.​10.​014 210–215. https://​doi.​org/​10.​1109/​KST.​2018.​84262​06
185. Shaw B, Suman AK, Chakraborty B (2020) Wine quality analy- 199. Mirzaee-Ghaleh E, Taheri-Garavand A, Ayari F, Lozano J
sis using machine learning. In J. K. Mandal & D. Bhattacharya (2020) Identification of fresh-chilled and frozen-thawed chicken
(Eds.), Emerging Technology in Modelling and Graphics 239– meat and estimation of their shelf life using an E-nose machine
247. Springer Singapore. coupled fuzzy KNN. Food Anal Methods 13(3):678–689.
186. Asnaashari M, Farhoosh R, Farahmandfar R (2016) Prediction https://​doi.​org/​10.​1007/​s12161-​019-​01682-6
of oxidation parameters of purified Kilka fish oil including gal- 200. Ayari F, Mirzaee- Ghaleh E, Rabbani H, Heidarbeigi K (2018)
lic acid and methyl gallate by adaptive neuro-fuzzy inference Using an E-nose machine for detection the adulteration of mar-
system (ANFIS) and artificial neural network. J Sci Food Agric garine in cow ghee. J Food Process Eng 41(6). https://​doi.o​ rg/​
96(13):4594–4602. https://​doi.​org/​10.​1002/​jsfa.​7677 10.​1111/​jfpe.​12806
187. Bahram-Parvar M, Salehi F, Razavi SMA (2017) Adaptive 201. Karami H, Rasekh M, Mirzaee-Ghaleh E (2020) Application
neuro-fuzzy inference system (ANFIS) simulation for predict- of the E-nose machine system to detect adulterations in mixed
ing overall acceptability of ice cream. EAEF 10(2):79–86. edible oils using chemometrics methods. J Food Process Preserv
https://​doi.​org/​10.​1016/j.​eaef.​2016.​11.​001 44(9):1–12. https://​doi.​org/​10.​1111/​jfpp.​14696
188. Neethu KC, Sharma AK, Pushpadass HA, Emerald FME, 202. Vajdi M, Varidi MJ, Varidi M, Mohebbi M (2019) Using elec-
Manjunatha M (2016) Prediction of convective heat transfer tronic nose to recognize fish spoilage with an optimum classifier.
coefficient during deep-fat frying of pantoa using neurocom- Journal of Food Measurement and Characterization 13(2):1205–
puting approaches. Innov Food Sci Emerg Technol 34:275– 1217. https://​doi.​org/​10.​1007/​s11694-​019-​00036-4
284. https://​doi.​org/​10.​1016/j.​ifset.​2016.​02.​012 203. Uçar A, Özalp R (2017) Efficient android electronic nose design
189. Mokarram M, Amin H, Khosravi MR (2019) Using adaptive for recognition and perception of fruit odors using Kernel Extreme
neuro-fuzzy inference system and multiple linear regression to Learning Machines. Chemom Intell Lab Syst 166:69–80. https://​
estimate orange taste. Food Sci Nutr 7(10):3176–3184. https://​ doi.​org/​10.​1016/j.​chemo​lab.​2017.​05.​013
doi.​org/​10.​1002/​fsn3.​1149
Food Engineering Reviews (2022) 14:134–175 173

204. Adak MF, Yumusak, N (2016) Classification of E-nose aroma ARTMAP) analysis of the data obtained with an electronic
data of four fruit types by ABC-based neural network. Sensors tongue applied to a ham-curing process with different salt for-
16(3). https://​doi.​org/​10.​3390/​s1603​0304 mulations. Applied Soft Computing Journal 30:421–429. https://​
205. Qiu S, Wang J (2017) The prediction of food additives in the fruit doi.​org/​10.​1016/j.​asoc.​2014.​12.​037
juice based on electronic nose with chemometrics. Food Chem 221. Marisol IJ-B, Luiz G-S, Ana P-M, Escriche. (2017) Antioxidant
230:208–214. https://​doi.​org/​10.​1016/j.​foodc​hem.​2017.​03.​011 activity and physicochemical parameters for the differentiation of
206. Faal S, Loghavi M, Kamgar S (2019) Physicochemical properties honey using a potentiometric electronic tongue. J Sci Food Agr
of Iranian ziziphus honey and emerging approach for predict- 97(7):2215–2222. https://​doi.​org/​10.​1002/​jsfa.​8031
ing them using electronic nose. Meas.: J Int Meas Confed 148, 222. Jingjing HL, Mingxu Z, Low Sze Shin Xu Ning, Chen Zhiqing
106936. https://​doi.​org/​10.​1016/j.​measu​rement.​2019.​106936 LV, Chuang Cui Ying, Shi Yan and Men (2020) Fuzzy evalua-
207. Guo T, Yin T, Ma Z, Wang Z, Sun X, Yuan W (2018) Characteri- tion output of taste information for liquor using electronic tongue
zation of different processes lemon slice using electronic tongue. based on cloud model. Sensors 20(3):1–20. https://​doi.​org/​10.​
IFAC-PapersOnLine 51(17):683–688. https://​doi.​org/​10.​1016/j.​ 3390/​s2003​0686
ifacol.​2018.​08.​117 223. Tohidi M, Ghasemi-Varnamkhasti M, Ghafarinia V, Bonyadian M,
208. Li J, Li Z, Li L, Song C, Raghavan GSV, He F (2021) Microwave Mohtasebi SS (2018) Development of a metal oxide semiconductor-
drying of balsam pear with online aroma detection and control. based artificial nose as a fast, reliable and non-expensive analyti-
J Food Eng 288(November 2019):110139. https://​doi.​org/​10.​ cal technique for aroma profiling of milk adulteration. Int Dairy J
1016/j.​jfood​eng.​2020.​110139 77:38–46. https://​doi.​org/​10.​1016/j.​idair​yj.​2017.​09.​003
209. Górska-Horczyczak E, Horczyczak M, Guzek D, Wojtasik- 224. Wang L, Niu Q, Hui Y, Jin H (2015) Discrimination of rice with
Kalinowska I, Wierzbicka A (2016) Chromatographic finger- different pretreatment methods by using a voltammetric elec-
prints supported by artificial neural network for differentiation tronic tongue. Sensors 15(7):17767–17785. https://​doi.​org/​10.​
of fresh and frozen pork. Food Control 73:237–244. https://​doi.​ 3390/​s1507​17767
org/​10.​1016/j.​foodc​ont.​2016.​08.​010 225. Wang L, Niu Q, Hui Y, Jin H, Chen S (2015) Assessment of
210. Srivastava S, Mishra G, Mishra HN (2019) Fuzzy controller taste attributes of peanut meal enzymatic-hydrolysis hydrolysates
based E-nose classification of Sitophilus oryzae infestation in using an electronic tongue. Sensors (Switzerland) 15(5):11169–
stored rice grain. Food Chem 283(January):604–610. https://d​ oi.​ 11188. https://​doi.​org/​10.​3390/​s1505​11169
org/​10.​1016/j.​foodc​hem.​2019.​01.​076 226. Hasan MA, Sarno R, Sabilla SI (2020) Optimizing machine
211. Srivastava S, Mishra G, Mishra HN (2019) Probabilistic artificial learning parameters for classifying the sweetness of pineapple
neural network and E-nose based classification of Rhyzopertha aroma using electronic nose. Int J Intell Syst 13(5):122–132.
dominica infestation in stored rice grains. Chemom Intell Lab https://​doi.​org/​10.​22266/​ijies​2020.​1031.​12
Syst 186:12–22. https://d​ oi.o​ rg/1​ 0.1​ 016/j.c​ hemol​ ab.2​ 019.0​ 1.0​ 07 227. Lu L, Tian S, Deng S, Zhu Z, Hu X (2015) Determination of rice
212. Raigar RK, Upadhyay R, Mishra HN (2017) Storage quality assess- sensory quality with similarity analysis-artificial neural network
ment of shelled peanuts using non-destructive electronic nose method in electronic tongue system. RSC Adv 5(59):47900–
combined with fuzzy logic approach. Postharvest Biol Technol 47908. https://​doi.​org/​10.​1039/​c5ra0​6310h
132(May):43–50. https://​doi.​org/​10.​1016/j.​posth​arvbio.​2017.​05.​016 228. De Sá AC, Cipri A, González-Calabuig A, Stradiotto NR, Del
213. Feng H, Zhang M, Liu P, Liu Y, Zhang X (2020) Evaluation of Valle M (2016) Resolution of galactose, glucose, xylose and
IoT-enabled monitoring and electronic nose spoilage detection for mannose in sugarcane bagasse employing a voltammetric elec-
salmon freshness during cold storage. Foods 9(11):1579. https://​doi.​ tronic tongue formed by metals oxy-hydroxide/MWCNT modi-
org/​10.​3390/​foods​91115​79 fied electrodes. Sens Actuators, B Chem 222:645–653. https://​
214. Heidarbeigi K, Mohtasebi SS, Foroughirad A, Ghasemi- doi.​org/​10.​1016/j.​snb.​2015.​08.​088
Varnamkhasti M, Rafiee S, Rezaei K (2015) Detection of adul- 229. Shi C, Yang X, Han S, Fan B, Zhao Z, Wu X, Qian J (2018)
teration in saffron samples using electronic nose. Int J Food Nondestructive prediction of tilapia fillet freshness during stor-
Prop 18(7):1391–1401. https://​doi.​org/​10.​1080/​10942​912.​2014.​ age at different temperatures by integrating an electronic nose
915850 and tongue with radial basis function neural networks. Food
215. Kiani S, Minaei S, Ghasemi-Varnamkhasti M (2016) A portable elec- Bioprocess Technol 11(10):1840–1852. https://​doi.​org/​10.​1007/​
tronic nose as an expert system for aroma-based classification of saf- s11947-​018-​2148-8
fron. Chemom Intell Lab Syst 156:148–156. https://​doi.​org/​10.​1016/j.​ 230. Shi Q, Guo T, Yin T, Wang Z, Li C, Sun X, Guo Y, Yuan W (2018)
chemo​lab.​2016.​05.​013 Classification of Pericarpium Citri Reticulatae of different ages by
216. Huang D, Bian Z, Qiu Q, Wang Y, Fan D, Wang X (2019) Identi- using a voltammetric electronic tongue system. Int J Electrochem
fication of similar Chinese Congou black teas using an electronic Sci 13(12):11359–11374. https://​doi.​org/​10.​20964/​2018.​12.​45
tongue combined with pattern recognition. Molecules 24(24):1– 231. Huang X, Yu S, Xu H, Aheto JH, Bonah E, Ma M, Wu M, Zhang
16. https://​doi.​org/​10.​3390/​molec​ules2​42445​49 X (2019) Rapid and nondestructive detection of freshness quality
217. Qiu S, Gao L, Wang J (2015) Classification and regression of of postharvest spinaches based on machine vision and electronic
ELM, LVQ and SVM for E-nose data of strawberry juice. J Food nose. J Food Saf 39(6):1–8. https://​doi.​org/​10.​1111/​jfs.​12708
Eng  144, 77–85. https://d​ oi.o​ rg/1​ 0.1​ 016/j.j​ foode​ ng.2​ 014.0​ 7.0​ 15 232. Fan S, Li J, Zhang Y, Tian X, Wang Q, He X, Zhang C, Huang
218. Upadhyay R, Sehwag S, Mishra HN (2017) Electronic nose W (2020) On line detection of defective apples using computer
guided determination of frying disposal time of sunflower oil vision system combined with deep learning methods. J Food Eng
using fuzzy logic analysis. Food Chem 221:379–385. https://d​ oi.​ 286:110102. https://​doi.​org/​10.​1016/j.​jfood​eng.​2020.​110102
org/​10.​1016/j.​foodc​hem.​2016.​10.​089 233. Nadian MH, Rafiee S, Aghbashlo M, Hosseinpour S, Mohtasebi
219. Mishra G, Srivastava S, Panda BK, Mishra HN (2018) Prediction SS (2015) Continuous real-time monitoring and neural network
of Sitophilus granarius infestation in stored wheat grain using mul- modeling of apple slices color changes during hot air drying.
tivariate chemometrics & fuzzy logic-based electronic nose analy- Food Bioprod Process 94:263–274. https://d​ oi.o​ rg/1​ 0.1​ 016/j.f​ bp.​
sis. Comput Electron Agric 152(July):324–332. https://​doi.​org/​10.​ 2014.​03.​005
1016/j.​compag.​2018.​07.​022 234. Mazen FMA, Nashat AA (2019) Ripeness classification of
220. Gil-Sánchez L, Garrigues J, Garcia-Breijo E, Grau R, Aliño M, bananas using an artificial neural network. Arab J Sci Eng
Baigts D, Barat JM (2015) Artificial neural networks (Fuzzy 44(8):6901–6910. https://​doi.​org/​10.​1007/​s13369-​018-​03695-5
174 Food Engineering Reviews (2022) 14:134–175

235. Lopes JF, Ludwig L, Barbin DF, Grossmann MVE, Barbon 250. Zareiforoush H, Minaei S, Alizadeh MR, Banakar A (2015) A
S (2019) Computer vision classification of barley flour based hybrid intelligent approach based on computer vision and fuzzy
on spatial pyramid partition ensemble. Sensors (Switzerland) logic for quality measurement of milled rice. Meas.: J Int Meas
19(13):1–17. https://​doi.​org/​10.​3390/​s1913​2953 Confed 66,26–34. https://​doi.​org/​10.​1016/j.​measu​rement.​2015.​
236. Gonzalez Viejo C, Torrico DD, Dunshea FR, Fuentes S (2019) 01.​022
Development of artificial neural network models to assess beer 251. Bakhshipour A, Zareiforoush H, Bagheri I (2020) Application
acceptability based on sensory properties using a robotic pourer: a of decision trees and fuzzy inference system for quality clas-
comparative model approach to achieve an artificial intelligence sys- sification and modeling of black and green tea based on visual
tem. Beverages 5(2):33. https://​doi.​org/​10.​3390/​bever​ages5​020033 features. Journal of Food Measurement and Characterization
237. Villaseñor-Aguilar MJ, Bravo-Sánchez MG, Padilla-Medina JA, 14(3):1402–1416. https://​doi.​org/​10.​1007/​s11694-​020-​00390-8
Vázquez-Vera JL, Guevara-González RG, García-Rodríguez FJ, 252. Wan P, Toudeshki A, Tan H, Ehsani R (2018) A methodology for
Barranco-Gutiérrez AI (2020) A maturity estimation of bell pep- fresh tomato maturity detection using computer vision. Comput
per (Capsicum annuum L.) by artificial vision system for quality Electron Agric 146(February 2017) 43–50. https://​doi.​org/​10.​
control. Appl Sci (Switzerland) 10(15):1–18. https://​doi.​org/​10.​ 1016/j.​compag.​2018.​01.​011
3390/​app10​155097 253. Garcia JAA, Arboleda ER, Galas EM (2020) Identification of
238. Castro W, Oblitas J, De-La-Torre M, Cotrina C, Bazan K, Avila- visually similar vegetable seeds using image processing and
George H (2019) Classification of cape gooseberry fruit accord- fuzzy logic. Int J Sci Technol Res 9(2):4925–4928
ing to its level of ripeness using machine learning techniques and 254. Curto B, Moreno V, García-Esteban JA, Blanco FJ, González I,
different color spaces. IEEE Access 7:27389–27400. https://​doi.​ Vivar A, Revilla I (2020) Accurate prediction of sensory attrib-
org/​10.​1109/​ACCESS.​2019.​28982​23 utes of cheese using near-infrared spectroscopy based on artifi-
239. Koklu M, Ozkan IA (2020) Multiclass classification of dry beans cial neural network. Sensors (Switzerland) 20(12):1–16. https://​
using computer vision and machine learning techniques. Comput doi.​org/​10.​3390/​s2012​3566
Electron Agric 174(June 2019):105507. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​ 255. Barbon S, Da Costa Barbon APA, Mantovani RG, Barbin DF
compag.​2020.​105507 (2018) Machine learning applied to near-infrared spectra for
240. Siswantoro J, Hilman MY, Widiasri M (2017) Computer vision chicken meat classification. J Spectrosc 2018:1–12. https://​doi.​
system for egg volume prediction using backpropagation neural org/​10.​1155/​2018/​89497​41
network. IOP Conference Series: Materials Science and Engi- 256. Gunaratne TM, Viejo CG, Gunaratne NM, Torrico DD, Dunshea
neering 273:2–7. https://​doi.​org/​10.​1088/​1757-​899x/​273/1/​ FR, Fuentes S (2019) Chocolate quality assessment based on chemi-
012002 cal fingerprinting using near infra-red and machine learning mod-
241. Khodaei J, Behroozi-khazaei N (2016) Combined application of eling. Foods 8(10):1–11. https://​doi.​org/​10.​3390/​foods​81004​26
decision tree and fuzzy logic techniques for intelligent grading 257. Arboleda ER (2018) Discrimination of civet coffee using near
of dried figs Food Process Engineering 1–12 https://​doi.​org/​10.​ infrared spectroscopy and artificial neural network. Int J Adv
1111/​jfpe.​12456 Comput Sci Appl 8(39):324–334. https://​doi.​org/​10.​19101/​
242. Huang X, Xu H, Wu L, Dai H, Yao L, Han F (2016) A data fusion IJACR.​2018.​839007
detection method for fish freshness based on computer vision 258. Zhang H, Sun H, Wang L, Wang S, Zhang W, Hu J (2018) Near
and near-infrared spectroscopy. Anal Methods 8(14):2929–2935. infrared spectroscopy based on supervised pattern recognition
https://​doi.​org/​10.​1039/​c5ay0​3005f methods for rapid identification of adulterated edible gelatin J
243. Rezagholi F, Hesarinejad MA (2017) Integration of fuzzy logic and Spectrosc 1–9 https://​doi.​org/​10.​1155/​2018/​76525​92
computer vision in intelligent quality control of celiac-friendly prod- 259. Aboonajmi M, Saberi A, Najafabadi TA, Kondo N (2016) Qual-
ucts. Procedia Computer Science 120:325–332. https://​doi.​org/​10.​ ity assessment of poultry egg based on visible-near infrared
1016/j.​procs.​2017.​11.​246 spectroscopy and radial basis function networks. Int J Food Prop
244. Phate VR, Malmathanraj R, Palanisamy P (2019) Clustered ANFIS 19(5):1163–1172. https://​doi.​org/​10.​1080/​10942​912.​2015.​10752​15
weighing models for sweet lime (Citrus limetta) using computer 260. Mohamed MY, Solihin MI, Astuti W, Ang CK, Zailah W (2019)
vision system. J Food Process Eng 42(6):1–16. https://​doi.​org/​10.​ Food powders classification using handheld Near-Infrared Spec-
1111/​jfpe.​13160 troscopy and Support Vector Machine. J Phys Conf Ser 1367(1).
245. Utai K, Nagle M, Hämmerle S, Spreer W, Mahayothee B, Müller https://​doi.​org/​10.​1088/​1742-​6596/​1367/1/​012029
J (2019) Mass estimation of mango fruits (Mangifera indica L., 261. Ren G, Wang Y, Ning J, Zhang Z (2020) Highly identification
cv. ‘Nam Dokmai’) by linking image processing and artificial of keemun black tea rank based on cognitive spectroscopy: near
neural network. Eng Agric Environ Food 12(1):103–110. https://​ infrared spectroscopy combined with feature variable selection.
doi.​org/​10.​1016/j.​eaef.​2018.​10.​003 Spectrochimica Acta - Part A: Molecular and Biomolecular Spec-
246. Nadim M, Ahmadifar H, Mashkinmojeh M, Yamaghani MR troscopy 230:118079. https://​doi.​org/​10.​1016/j.​saa.​2020.​118079
(2019) Application of image processing techniques for quality 262. Alshejari A, Kodogiannis VS (2017) An intelligent decision sup-
control of mushroom. Caspian J Health Res 4(3):72–75. https://​ port system for the detection of meat spoilage using multispectral
doi.​org/​10.​29252/​cjhr.4.​3.​72 images. Neural Comput Appl 28(12):3903–3920. https://d​ oi.o​ rg/​
247. Sidehabi SW, Suyuti A, Areni IS, Nurtanio I (2018) The devel- 10.​1007/​s00521-​016-​2296-6
opment of machine vision system for sorting passion fruit 263. Xu Q, Yang X, Hu Z, Li Y, Zheng X (2020) Identification of
using Multi-Class Support Vector Machine. J Eng Sci Tech- adulterated olive oil by fusion of near infrared and Raman spec-
nol 11(5):178–184. https://​doi.​org/​10.​25103/​jestr.​115.​23 troscopy. J Phys: Conf Ser 1592:1–7. https://​doi.​org/​10.​1088/​
248. Sun X, Young J, Liu JH, Newman D (2018) Prediction of pork 1742-​6596/​1592/1/​012041
loin quality using online computer vision system and artificial 264. Lu M, Li C, Li L, Wu Y, Yang Y (2018) Rapid detecting soluble
intelligence model. Meat  Sci  140(November 2017):72–77. solid content of pears based on near-infrared spectroscopy. Pro-
https://​doi.​org/​10.​1016/j.​meats​ci.​2018.​03.​005 ceedings of 2018 2nd IEEE Advanced Information Management,
249. Bhagat NB, Markande SD (2017) Automatic grading of potatoes Communicates, Electronic and Automation Control Conference,
with fuzzy logic. Proceedings of 2016 Online International Con- IMCEC 2018, Imcec, 819–823. https://d​ oi.o​ rg/1​ 0.1​ 109/​IMCEC.​
ference on Green Engineering and Technologies, IC-GET 2016. 2018.​84693​15
https://​doi.​org/​10.​1109/​GET.​2016.​79166​51
Food Engineering Reviews (2022) 14:134–175 175

265. Rizwana S, Hazarika MK (2020) Application of near-infrared 267. Richter B, Rurik M, Gurk S, Kohlbacher O, Fischer M (2019)
spectroscopy for rice characterization using machine learn- Food monitoring: screening of the geographical origin of white
ing. J Inst Eng (India): A 101(4):579–587. https://​doi.​org/​10.​ asparagus using FT-NIR and machine learning. Food Control
1007/​s40030-​020-​00459-z 104(February):318–325. https://d​ oi.o​ rg/1​ 0.1​ 016/j.f​ oodco​ nt.2​ 019.​
266. Barbon Junior S, Mastelini SM, Barbon APAC, Barbin DF, 04.​032
Calvini R, Lopes JF, Ulrici A (2020) Multi-target prediction of
wheat flour quality parameters with near infrared spectroscopy. Publisher’s Note Springer Nature remains neutral with regard to
Information Processing in Agriculture 7(2):342–354. https://​doi.​ jurisdictional claims in published maps and institutional affiliations.
org/​10.​1016/j.​inpa.​2019.​07.​001