sustainability

Article
Fruit Image Classification Model Based on MobileNetV2 with
Deep Transfer Learning Technique
Yonis Gulzar

Department of Management Information Systems, College of Business Administration, King Faisal University,
Al-Ahsa 31982, Saudi Arabia; [email protected]; Tel.: +966-545-719-118

Abstract: Due to the rapid emergence and evolution of AI applications, the utilization of smart
imaging devices has increased significantly. Researchers have started using deep learning models,
such as CNN, for image classification. Unlike the traditional models, which require a lot of features
to perform well, CNN does not require any handcrafted features to perform well. It uses numerous
filters, which extract required features from images automatically for classification. One of the issues
in the horticulture industry is fruit classification, which requires an expert with a lot of experience.
To overcome this issue an automated system is required which can classify different types of fruits
without the need for any human effort. In this study, a dataset of a total of 26,149 images of 40 different
types of fruits was used for experimentation. The training and test set were randomly recreated and
divided into the ratio of 3:1. The experiment introduces a customized head of five different layers
into MobileNetV2 architecture. The classification layer of the MobileNetV2 model is replaced by the
customized head, which produced the modified version of MobileNetV2 called TL-MobileNetV2.
In addition, transfer learning is used to retain the pre-trained model. TL-MobileNetV2 achieves an
accuracy of 99%, which is 3% higher than MobileNetV2, and the equal error rate of TL-MobileNetV2
is just 1%. Compared to AlexNet, VGG16, InceptionV3, and ResNet, the accuracy is better by 8, 11, 6,
and 10%, respectively. Furthermore, the TL-MobileNetV2 model obtained 99% precision, 99% for
recall, and a 99% F1-score. It can be concluded that transfer learning plays a big part in achieving
better results, and the dropout technique helps to reduce the overfitting in transfer learning.

Keywords: deep learning; classification; fruits; MobileNetV2; precision agriculture

Citation: Gulzar, Y. Fruit Image
Classification Model Based on
MobileNetV2 with Deep Transfer
Learning Technique. Sustainability 1. Introduction
2023, 15, 1906. https://doi.org/ Artificial intelligence is becoming a driving force in all aspects of life, and the agri-
10.3390/su15031906 cultural and food industries are not far behind. Artificial intelligence has been used in
Academic Editor: Gwanggil Jeon
numerous areas, such as healthcare, education, agriculture, and many other areas. In the
area of healthcare, AI has been used for diagnosing numerous diseases, such as skin cancer
Received: 21 December 2022 disease [1], identifying different anatomy objects [2], predicting neurodevelopmental dis-
Revised: 9 January 2023 orders in children [3], mental health, and other issues [4,5]. When it comes to agriculture,
Accepted: 17 January 2023 the world is facing challenges, such as increasing global population, global warming, and
Published: 19 January 2023
other human-caused environmental hazards, which may eventually lead to an increase
in demand for food supplies. This is where AI and the computer-vision-driven Agtech
industry appears to come to the rescue by speeding up the process of harvesting, quality
control, picking and packing, sorting, grading, and other processes [6]. When it comes
Copyright: © 2023 by the author.
Licensee MDPI, Basel, Switzerland.
to fruits, they are very delicate and decay quickly. Around 30–35% of harvested fruits
This article is an open access article
get wasted due to improper and delayed identification, classification, and the grading
distributed under the terms and of fruits undertaken by non-skilled workers. Fruit classification is considered the most
conditions of the Creative Commons difficult and vital process when it comes to selling/purchasing fruit. A person dealing
Attribution (CC BY) license (https:// with selling/buying fruits needs to have knowledge of the different varieties of fruit(s) for
creativecommons.org/licenses/by/ prizing purposes. So, a person needs to have good knowledge of recognizing the different
4.0/). varieties of fruits.

Sustainability 2023, 15, 1906. https://doi.org/10.3390/su15031906 https://www.mdpi.com/journal/sustainability

Sustainability 2023, 15, 1906 2 of 14

Many fruit, vegetable, and seed identification, classification, and grading methods
have been developed [7]. There have been different classification methods proposed for
different classes of fruits. For instance, Altaheri et al. [8] proposed a robotic harvesting
model designed to classify five different types of date fruits. This model achieved around
99% accuracy. This model used an in-house dataset for training and testing. The dataset
contained overall 8000 images. In another study, Shamim Hossain et al. [9] developed
a fruit classification model for industrial applications. They used the publicly available
dataset to train and test their model. One of the datasets contained images of fruits which
are complex to identify. The proposed model achieved an accuracy of 85%. Gulzar et al. [10]
proposed a model for seed classification based on VGG16. They used thirteen types of
different seeds, and the model achieved 99% accuracy. On the other hand, Hamid et al. [11]
proposed a model based on the same dataset and used MobileNetV2 as the base model.
They incorporated a transfer learning technique [12], and the model acquired 99% accuracy.
Saranya et al. [13] undertook a comparative study in which they trained different machine
learning and deep learning models on a public dataset. This dataset contains images of
different fruits, such as apples, bananas, oranges, and pomegranates. They concluded that
deep learning-based models outperform machine learning models. Rojas-Aranda et al. [14]
developed a model to classify fruits in retail stores using deep learning. The purpose of this
study was to improve the checkout process in retail stores. The model showed an accuracy
of 95% when the fruits were within the plastic bags, whereas the accuracy was recorded
as 93% when the fruits were not covered by plastic. Sridhar et al. [15] proposed a model
for 31 different types of fruits using a hybrid approach. They incorporated CNN and an
autoencoder to handle these huge data of 31 different fruits. They claim that their model
achieved 99% of accuracy. Zhou et al. [16] developed a model to detect the plumpness of the
strawberry fruit. They attained around 86% accuracy in terms of detecting strawberries in
the greenhouse. They used RGD data while training the proposed model. Mamat et al. [17]
proposed a model based on deep learning using with Only Look Once (YOLO) versions
and adopted transfer learning for palm oil fruit. The model attained 98.7% accuracy for
palm oil fruit.
Some researchers focused on the identification and classification of fruit diseases [18].
This study used VGG19 architecture as a base model. They claimed that their proposed
model obtained around 99% accuracy in classifying fruits and their diseases. In another
study, Assuncao et al. [19] proposed a deep learning model to operate on mobile devices.
This model aims to classify peaches based on their freshness as well as to identify three
types of diseases found in peach fruit. The accuracy of the model was recorded as 96%.
They incorporated some preprocessing techniques to improve the accuracy of the pro-
posed model.
There have been some studies that focused on the quality of the fruits, such as [20–23].
Garillos-Manliguez et al. [20] proposed a model for the estimation of the maturity of
papaya fruit. The unique thing about this model is that it is trained on hyperspectral
and visible-light images, unlike other models. These images not only show the external
characteristics but also provide details about the inside of the fruit. The model acquired 97%
of accuracy in terms of estimating the maturity of papaya fruit. Herman et al. [21] chose oil
palm fruit to check its ripeness. The dataset they used contained around seven different
types of ripeness levels of the oil palm fruit. They trained two well-known architectures
(AlexNet and DenseNet) on this dataset and concluded that DenseNet outperformed
AlexNet in terms of accuracy by 8%. Mahmood et al. [23] performed a comparative study
on two well-known architectures (AlexNet and VGG16) to check the maturity level of
jujube fruit. The dataset contained three different varieties of images in terms of ripeness
(unripe, ripe, and over-ripe). They also used some preprocessing techniques, such as
data augmentation. They claimed that VGG16 outperformed the AlexNet architecture by
achieving an accuracy of 98%.
Sustainability 2023, 15, 1906 3 of 14

When it comes to apple fruits, the apple is a Rosaceae family fruit that originated in
Asia. It is grown in over 63 countries throughout the world, with China being the main
producer. Due to their high-water content, carbohydrates, organic acids, vitamins, minerals,
and dietary fibers, apples are regarded as the most nutritious food. The apple, which is the
fourth most widely cultivated and consumed fruit on the planet, can be divided into several
types depending on its qualitative characteristics [24]. There are around 7500 different
varieties of apples found in the world [25]. Different apple varieties have different benefits
when it comes to health. For a common person, it is not easy to identify all kinds of apples
and other fruits with many varieties. Therefore, there is a need for an approach/model
based on deep learning, which can identify different kinds of fruits and solve the problem
of being dependent on an expert and improve the efficiency and accuracy in identifying
and classifying different fruit types.
In this study, a deep learning approach was proposed for the classification and identi-
fication of different kinds of fruits. The proposed model incorporates a transfer learning
technique, which helps to solve problems involving issues of insufficient training data.
This technique encourages the idea of not training the model from scratch and significantly
helps in reducing training time. In this study, a well-known deep learning model, Mo-
bileNetV2 [26], was used as the base model but was modified by adding five different
layers for improving the accuracy and reducing the error rate during the classification
process. The proposed model is trained on a dataset containing 40 varieties of fruits. The
results show that the proposed model achieved the highest accuracy rate in identifying
different types of fruits. The proposed model can be deployed in a mobile application for
practical usage. Further details about it are mentioned in Section 2.
The following points summarize the contributions of this paper:
• A detailed review was conducted to examine the most notable work in fruit classifica-
tion via machine learning and deep learning.
• A fruit classification problem was re-introduced based on a pre-trained MobileNetV2
CNN model, in which different kinds of fruits were classified.
• A modified model was proposed using advanced deep learning techniques for fruit
classification, and different model-tuning techniques were used to reduce the chances
of model overfitting, such as dropout and data augmentation techniques.
• An optimization technique was developed to monitor any positive change in the
validation accuracy and validation error rate. In case of change, a backup of an optimal
model was taken to make sure that the proposed model shows optimal accuracy and
the least validation loss.
The remainder of this article is organized as follows: In Section 2, the description of
the dataset, model selection, proposed model, model tuning, and experimental settings are
reported and discussed. In Section 3, the results and discussions are provided, whereas
Section 4 describes the conclusion.

2. Materials and Methods

2.1. Dataset
In this research work, a public dataset [27] was used, containing images of forty
different types of fruits, such as Braeburn, Crimson Snow, Golden 1, Golden 2, Golden 3,
Granny Smith, Pink Lady, Red 1, Red 2, Red 3, Red Delicious, Red Yellow 1, Red Yellow 2,
Apricot, Avocado, Banana, Banana Lady Finger, Banana Red, Beetroot, Blueberry, Cactus
Fruit, Cantaloupe 1, Cantaloupe 2, Carambola, Cherry, Cherry Wax Black, Cherry Wax Red,
Cherry Wax Yellow, Chestnut, Clementine, Cocos, Grape Blue, Grape White, Grapefruit
Pink, Grapefruit White, Guava, Huckleberry, Kaki, Kiwi, and Kumquats. The dataset is
divided into two subsets training, and testing datasets. The original number of images per
variety of the fruit is shown in Table 1.
 Sustainability 2023, 15, x FOR PEER REVIEW 4 of 15

Sustainability 2023, 15, 1906 4 of 14

Fruit Types No. of Images Fruit Types No. of Images
Apple Braeburn 640 Cactus Fruit 654
Apple
Table Crimson
1. Details Snow types of fruits
of different 604and their image
Cantaloupe
count. 1 658
Apple Golden 1 644 Cantaloupe 2 656
Fruit Types No. of Images Fruit Types No. of Images
Apple Golden 2 653 Carambola 738
Apple
AppleBraeburn
Golden 3 640
645 Cactus
CherryFruit 656654
Apple Crimson Snow 604 Cantaloupe 1 658
Apple Granny Smith
Apple Golden 1
646
644
Cherry Wax Black
Cantaloupe 2
656656
Apple Pink
Apple Golden 2 Lady 620
653 Cherry Wax Red
Carambola 645738
Apple
AppleGolden
Red 31 645
656 Cherry Cherry
Wax Yellow 658656
Apple Granny
Apple RedSmith
2 646
636 Cherry Wax Black
Chestnut 616656
Apple Pink Lady 620 Cherry Wax Red 645
AppleRed
Apple Red1 3 595
656 Clementine
Cherry Wax Yellow 818658
Apple Red Delicious
Apple Red 2 654
636 Cocos
Chestnut 654616
Apple
Apple RedRed 3
Yellow 1 595
711 Clementine
Grape Blue 656818
Apple Red Delicious 654 Cocos 654
Apple Red Yellow 2 836 Grape White 658
Apple Red Yellow 1 711 Grape Blue 656
Apricot
Apple Red Yellow 2 635
836 Grapefruit
Grape WhitePink 656658
Avocado
Apricot 593
635 Grapefruit
GrapefruitWhite
Pink 658656
Avocado
Banana 593
642 Grapefruit
Guava White 656658
Banana 642 Guava 656
Banana Lady Finger
Banana Lady Finger
616
616
Huckleberry
Huckleberry
656656
BananaRed
Banana Red 640
640 Kaki
Kaki 654654
Beetroot
Beetroot 604
604 Kiwi
Kiwi 622622
Blueberry
Blueberry 626
626 Kumquats
Kumquats 628628

The fruits were planted in a rotating motor shaft (3 rmp), and a short video of 20 s was
The fruits were planted in a rotating motor shaft (3 rmp), and a short video of 20 s
taken for each class. The images were taken from that video. The background of the images
was taken for each class. The images were taken from that video. The background of the
was removed with an algorithm because of variations in the lighting conditions. Figure 1
images was removed with an algorithm because of variations in the lighting conditions.
shows the sample captured images of fruits.
Figure 1 shows the sample captured images of fruits.

Figure 1. Sample images of forty types of fruits.

Figure 1. Sample images of forty types of fruits.
 Sustainability 2023, 15, x FOR PEER REVIEW 5 of 15

Sustainability 2023, 15, 1906 5 of 14

2.2. Model Selection

In the
2.2. Model field of image processing, CNN has gained more attention due to its huge eco-
Selection
nomic potential and high accuracy rate. There are well-known CNN architectures, which
In the field of image processing, CNN has gained more attention due to its huge
are very popular in the area of image processing and classification, such as AlexNet [28],
economic potential and high accuracy rate. There are well-known CNN architectures, which
VGG16 [29], InceptionV3 [30], ResNet [31], and MobileNetV2 [26]. In any computer vision
are very popular in the area of image processing and classification, such as AlexNet [28],
task, it is the convolution operation(s) which has a high contribution. Nevertheless, due
VGG16 [29], InceptionV3 [30], ResNet [31], and MobileNetV2 [26]. In any computer vision
to large and deep network structures, the processing time, as well as cost, increases in
task, it is the convolution operation(s) which has a high contribution. Nevertheless, due to
most of these networks, such as AlexNet, VGG16, InceptionV3, ResNet, etc. However,
large and deep network structures, the processing time, as well as cost, increases in most
when it comes to MobileNetV2, its inverse residual structure and linear bottleneck struc-
of these networks, such as AlexNet, VGG16, InceptionV3, ResNet, etc. However, when it
ture help to reduce the calculation of convolution. Its architecture is preferred over others
comes to MobileNetV2, its inverse residual structure and linear bottleneck structure help to
due to its simple architecture and memory-efficient characteristics. Table 2 presents the
reduce the calculation of convolution. Its architecture is preferred over others due to its
precision,
simple recall, and
architecture F1-score of AlexNet,
and memory-efficient VGG16, InceptionV3,
characteristics. Table 2 presents ResNet, and Mo-
the precision,
recall, and F1-score of AlexNet, VGG16, InceptionV3, ResNet, and MobileNetV2.dataset
bileNetV2. It is important to note that all the models were trained on the fruit It is
without using
important to noteany pre-processing
that all the modelstechniques.
were trained Only
on the
the classification layer was
fruit dataset without modified
using any
based on the techniques.
pre-processing number of Onlyclasses
thepresent in thelayer
classification said was
dataset. As mentioned
modified based on the before,
number Mo-
of classes present in the said dataset. As mentioned before, MobileNetV2 is memory-to
bileNetV2 is memory-efficient and was mainly designed for mobile applications. Due
this, it can
efficient andbe wasseen fromdesigned
mainly the tablefor
thatmobile
it is outperforming all oftothe
applications. Due other
this, models
it can in from
be seen terms
of accuracy.
the table that it is outperforming all of the other models in terms of accuracy.

Table
Table 2. 2. Precision,
Precision, recall,
recall, andand F1-score
F1-score ofof different
different models
models while
while trained
trained onon
thethe fruit
fruit dataset.
dataset.
Models Precision Recall F1-Score
Models Precision Recall F1-Score
AlexNet 0.81 0.79 0.83
AlexNet 0.81 0.79 0.83
VGG16
VGG16 0.79
0.79 0.80
0.80 0.80
0.80
InceptionV3
InceptionV3 0.84 0.84 0.82 0.82 0.810.81
ResNet ResNet 0.82 0.82 0.81 0.81 0.820.82
MobileNetV2
MobileNetV2 0.89 0.89 0.91 0.91 0.890.89

Figure
Figure2 2shows
showsthe
theaccuracy
accuracyofofdifferent
differentmodels
modelsachieved
achievedwhile
whiletraining
trainingon
onthe
thefruit
fruit
dataset containing forty different types of fruits. From the figure, it is evident that Mo-
dataset containing forty different types of fruits. From the figure, it is evident that Mo-
bileNetV2 achieves the highest accuracy with 89%, which is higher than ResNet, Incep-
bileNetV2 achieves the highest accuracy with 89%, which is higher than ResNet, Incep-
tionV3, VGG16, and AlexNet by 7, 5, 11, and 7%, respectively.
tionV3, VGG16, and AlexNet by 7, 5, 11, and 7%, respectively.

Figure 2. Accuracy rate of different models during model selection.

Figure 2. Accuracy rate of different models during model selection.

2.3.TL-MobileNetV2
2.3. TL-MobileNetV2
Thissubsection
This subsection focuses
focuses onon
thethe proposed
proposed model
model based
based onfruit
on the the fruit classification
classification da-
dataset.
taset.
The The proposed
proposed model, model, TL-MobileNetV2,
TL-MobileNetV2, is on
is based based on MobileNetV2
MobileNetV2 architecture
architecture [26].
[26]. The
Thearchitecture
said said architecture
is mainlyisdeveloped
mainly for
developed
mobile andforresource-constrained
mobile and resource-constrained
environments.
The main purpose for adopting this type of architecture lies in its strength in terms of
 ertheless, in order to make it matchable with the said problem, it is important to fix
number of nodes in the classification layer. To improve the accuracy, a customized h
containing five different layers was attached to the base model of MobileNetV2 by rep
ing the classification layer.
The customized head contains (i) average pooling layer, (ii) flatten layer, (iii) de
Sustainability 2023, 15, 1906 layer, (iv) dropout layer, and (v) softmax layer, as shown in Figure 3.6 Itof is 14 importan
mention that the size of the average pooling layer is set to (7 × 7). In the flatten layer
flattened neurons were fed to a dense layer with the activation function being ReLU.
was proceeded by setting the probability of a dropout layer with a value of 0.5 and
reducing the memory usage, computational expenses, and compliance of its design to be
addition of forty nodes within the classification layer of the model. Carrying out all th
more focused towards mobile applications. It is important to note that the said architecture
changes has produced an improved version of the MobileNetV2 architecture bearing f
has initially around one thousand nodes in its classification layer. Nevertheless, in order to
different nodes in its last (classification) layer; this is an optimal and befitting mode
make it matchable with the said problem, it is important to fix the number of nodes in the
the said problem addressed in this work.
classification layer. ToFrom
improve the accuracy, a customized head, containing five different
the literature, it is evident that transfer learning [32] has improved the per
layers was attached to the base model
mance of target learners ofon
MobileNetV2
target domains. by replacing
As a result,the classification
the dependencylayer.
on a large volu
The customized
of datahead
hascontains
decreased,(i)and
average pooling
the issues layer,
related to (ii)
the flatten layer, (iii)
unavailability dense data w
of sufficient
layer, (iv) dropout layer, [33,34].
reduced and (v)There
softmax
are layer,
different as shown
ways toin Figure transfer
integrate 3. It is important
learning intodeep learn
mention that the models;
size of the average pooling layer is set to (7 × 7). In the
one way is to train the model totally based on the new dataset.flatten layer,Another
the way
flattened neuronstrain
were fed to a dense layer with the activation function being ReLU.
only those layers of the model which were added to the existing model, whereas This
was proceeded by setting
existing the probability
layers are frozen. Inof this
a dropout
researchlayer
work,with a value
a hybrid of 0.5 and
approach was the
opted, wher
addition of forty the
nodesbeginning of training the model, only the newly added layers werethese
within the classification layer of the model. Carrying out all trained based
changes has produced an dataset.
the fruit improved version
After the 20ofth the MobileNetV2
iteration, architecture
the existing bearing
layers were forty and a sl
unfrozen,
different nodes inweight
its lastadjustment
(classification)
waslayer; this isto
performed anthe
optimal
trained and befitting
layers of themodel
modelfor the on the
based
said problem addressed
dataset. in this work.

Figure 3. TL-MobileNetV2 model.

Figure 3. TL-MobileNetV2 model.
2.4. ModelitTuning
From the literature, is evident that transfer learning [32] has improved the perfor-
mance of target learners on target domains. As a result, the dependency on a large volume
of data has decreased, and the issues related to the unavailability of sufficient data were
reduced [33,34]. There are different ways to integrate transfer learning in deep learning
models; one way is to train the model totally based on the new dataset. Another way is to
train only those layers of the model which were added to the existing model, whereas the
existing layers are frozen. In this research work, a hybrid approach was opted, where at the
beginning of training the model, only the newly added layers were trained based on the
fruit dataset. After the 20th iteration, the existing layers were unfrozen, and a slight weight
adjustment was performed to the trained layers of the model based on the said dataset.

2.4. Model Tuning

In this work, we have incorporated different preprocessing/model tuning techniques
to prevent the model from overfitting. The following are those techniques briefly explained:
• Data Augmentation: one of the best ways is to overcome the issue of insufficient data
is to use the data augmentation technique. This technique helps to create random
artificial images from the source data using different methods, such as shifts, shears,
 • Data Augmentation: one of the best ways is to overcome the issue of insufficient d
is to use the data augmentation technique. This technique helps to create rand
artificial images from the source data using different methods, such as shifts, she
random rotations, and flips. In this work, a built-in function in the Keras library
Sustainability 2023, 15, 1906 was used to create ten images from each image by randomly zooming7 of 20%,
14 adjus
height by 10%, width shifting by 10%, and rotating by 30%.
• Adaptive Learning Rate: this technique helps to accelerate the training and allevi
the pressure
random of choosing
rotations, and flips. In a this
learning
work, rate and function
a built-in schedule. In this
in the Keraswork,
librarythe
[35]initial le
was
ing used
ratetowas
createset
tentoimages from each
INIT_LR image by
= 0.001, andrandomly
then thezooming
decay 20%,of adjusting
the form deca
height by 10%, width shifting
INIT_LT/EPOCHS by 10%, and rotating by 30%.
was used.
• Adaptive Learning Rate: this technique helps to accelerate the training and alle-
• Model Checkpointing: while training the model, a checkpoint is set to monitor wh
viates the pressure of choosing a learning rate and schedule. In this work, the
ever there
initial is any
learning rate positive
was set to change
INIT_LR in accuracy.
= 0.001, and The model’s
then the decayweights
of theare
formsaved wh
ever the
decay accuracy is optimum.
= INIT_LT/EPOCHS was used. In this research work, a model checkpoint of the f
• checkpoint
Model = Model
Checkpointing: Checkpoint
while training the (fname, monitor is= set
model, a checkpoint “val_loss”, mode = “m
to monitor when-
ever there is any positive change in accuracy. The model’s weights
save_best_only = True, verbose = 1) was used. This callback monitors the valida are saved when-
ever the accuracy is optimum. In this research work, a model checkpoint of the
loss of the model and overwrites the trained model only when there is a decreas
form checkpoint = Model Checkpoint (fname, monitor = “val_loss”, mode = “min”,
the loss compared to the previous best model.
save_best_only = True, verbose = 1) was used. This callback monitors the validation
• Dropout:
loss The dropout
of the model technique
and overwrites helpsmodel
the trained avoidonly
the when
issue there
of overfitting,
is a decreaseand
in during
training, neurons are randomly
the loss compared to the previous best model.chosen and discarded. It implies that the contribu
• Dropout:
of these The dropout
neurons technique helps
is temporally avoid to
ignored thethe
issue of overfitting,
activation and during the
of downstream neurons,
training, neurons are randomly chosen and discarded. It implies
any weight changes are not implemented on any neuron on the backward pass. that the contribution
of these neurons is temporally ignored to the activation of downstream neurons, and
any weight changes are not implemented on any neuron on the backward pass.
2.5. Experimental Environment Settings and Model Evaluation Indicator
2.5. Experimental
This research Environment
aims toSettings
proposeandan
Model Evaluation
optimal modelIndicator
which identifies and classifies
ferentThis research
types aims to The
of fruits. propose an optimal
proposed modelwas
model which identifies and classifies
implemented different3.0 on W
using Python
dows 10 OS, with system configuration using an i7 processor and 16 GB 10
types of fruits. The proposed model was implemented using Python 3.0 on Windows OS,
RAM.
with system configuration using an i7 processor and 16 GB RAM.
The performance of the model is measured using a wide range of matrices deri
The performance of the model is measured using a wide range of matrices derived
from
from aa22××22confusion
confusion matrix,
matrix, as shown
as shown in Figure
in Figure 4. 4.

Figure
Figure 4.4.Confusion
Confusion matrix
matrix (Ci, j(C
). i, j).

When incorporating machine learning techniques, it is important to evaluate the model

When incorporating machine learning techniques, it is important to evaluate
to check its performance. For that reason, the matrices, as follows, were used to evaluate
model to check
the proposed its performance. For that reason, the matrices, as follows, were use
model:
evaluate
•
the proposed
Accuracy =
model:
TP+ TN
2TP+ FP+ TN + FNa
𝑇𝑃+𝑇𝑁
•• 𝑨𝒄𝒄𝒖𝒓𝒂𝒄𝒚
Precision = TP =TP
+2𝑇𝑃+𝐹𝑃+𝑇𝑁+𝐹𝑁𝑎
FP
TP 𝑇𝑃
•• Recall =
𝑷𝒓𝒆𝒄𝒊𝒔𝒊𝒐𝒏 =
TP + FN
• 𝑇𝑃+𝐹𝑃
Where TP, TN,𝑇𝑃FP, and FN are true positive, true negative, false positive, and false
• 𝑹𝒆𝒄𝒂𝒍𝒍 =
negative, respectively
𝑇𝑃+𝐹𝑁
• Macro Average: The function to compute F-1 for each label and returns the average
without considering the proportion for each label in the dataset.
• Weighted Average: The function to compute F-1 for each label and returns the average
considering the proportion for each label in the dataset.
 continuous monitoring is used to spot any noticeable deviation in training and validatio
performance when it comes to validation accuracy and loss.

3. Results
Sustainability 2023, 15, 1906 This section presents the performance of the proposed model in terms of trainin
8 of 14
accuracy, training loss, as well as validation accuracy, as shown in Figure 5. The propose
model was trained for 100 iterations. From the figure, it can be noticed that the trainin
accuracy of themodel
The trained proposed model
is further started
assessed withvalidation
using 44% fromaccuracythe firstand
iteration, and the accurac
loss. Moreover,
increasedmonitoring
continuous dramatically. Thetoaccuracy
is used touched deviation
spot any noticeable 90% within the first
in training and10validation
iterations. At th
performance when
30th iteration, it model
the comes to validation
reached accuracy accuracy
maximum and loss. (100%). From the 30th iteration on
wards, it can be inferred that the accuracy rate of the proposed model has remained at th
3. Results
maximum until the end of the training.
This section presents the performance of the proposed model in terms of training
Figure 5 also shows the training loss of the proposed model. From the figure, it
accuracy, training loss, as well as validation accuracy, as shown in Figure 5. The proposed
depicted
model was that at the
trained for beginning of the
100 iterations. Fromtraining phase,
the figure, thebetraining
it can noticed loss is high
that the as the mod
training
has not been exposed to the data. However, gradually the model
accuracy of the proposed model started with 44% from the first iteration, and the accuracy reads the images an
starts todramatically.
increased remember them, and eventually
The accuracy touched 90% thewithin
training loss 10
the first gets reduced.
iterations. At It
thecan
30th be notice
that thethe
iteration, training loss reaches
model reached 0.6 within
maximum accuracythe first 20
(100%). Fromiterations and getsonwards,
the 30th iteration reducedit dramat
can be inferred
cally by eachthat the accuracy
iteration. By therate of the
100th proposed
iteration, themodel has remained
training at the maximum
loss has reached 0.3 which infer
until the end of the training.
the characteristics of a good model.

Figure
Figure 5. 5. Training
Training accuracy
accuracy and and loss
loss in in terms
terms of training
of training and validation
and validation of TL-MobileNetV2
of TL-MobileNetV2 model. mode

Figure 5 also shows

Generally, most ofthethe
training
models lossperform
of the proposed model.
well during From the
training. figure, it is
However, they don
depicted that at the beginning of the training phase, the training loss is high as the model
perform well during the validation phase. This is due to the fact that the model has onl
has not been exposed to the data. However, gradually the model reads the images and
beentotrained
starts rememberon them,
supervised data. In the
and eventually order to find
training lossout
getsitsreduced.
performance,
It can beitnoticed
is important t
validate
that the performance
the training ofwithin
loss reaches 0.6 the model.
the firstFor this purpose,
20 iterations and gets thereduced
proposed model was val
dramatically
by each iteration. By the 100th iteration, the training loss has reached 0.3 which infers the model.
dated as well. Figure 5 shows the validation accuracy and loss of the proposed
characteristics of a good model.
Generally, most of the models perform well during training. However, they don’t
perform well during the validation phase. This is due to the fact that the model has only
been trained on supervised data. In order to find out its performance, it is important
to validate the performance of the model. For this purpose, the proposed model was
validated as well. Figure 5 shows the validation accuracy and loss of the proposed model.
It depicts that the model starts with 15% of accuracy at the beginning of the process and
within 10 iterations, the model reaches 100%. From the 5th iteration onwards, the model’s
validation accuracy remains steady. When it comes to the validation loss of the proposed
model, it can be seen from the figure that the model starts with a very high validation loss,
Sustainability 2023, 15, 1906 9 of 14

which is common in all deep learning models. The validation loss of the proposed model
dramatically falls and reaches 0.5 within the first 10 iterations. From the 10th iteration
onwards until the 50th iteration, the validation loss remains constant and then starts falling
again from the 55th iteration. Finally, it reaches 0.35 at the end of the 100th iteration.
During the training and validation process, the model shows stability in its perfor-
mance. The proposed model shows a very high accuracy rate during training, and it
reflected the same during validation. When it comes to training loss and validation loss, the
model proves that it does not overfit. This is due to the fact the preprocessing techniques
used in the proposed model helped the model to achieve better results without overfitting.
Moreover, the data augmentation technique incorporated in the proposed model played a
vital role in exposing the model to different variations of the images. Additionally, using
of dropout technique helped the model’s validation performance by making sure that the
model does not deviate much from its training performance.
Table 3 shows the performance of the TL-MobileNetV2 model based on each fruit
class in terms of precision, recall, F1-score, and support. The supports show the number of
images used for training and validating the model after applying the data augmentation
technique. As mentioned earlier, for each instance of the image, the image augmentation
technique creates ten different instances of that image, which are used for model training.
It can be inferred from the table that the model achieved the maximum value for each class
when it comes to precision, recall, and F1-score except for some of fruits, such as Apple
Golden 1, Apple Red 3, Apple Red Yellow 2, Banana Lady Finger, Cantaloupe 2, Cherry
Wax Red, and Kaki. This is due to the fact that the dataset does not have any variations
present in the background of any images. As stated earlier, the background of each instance
was removed, which makes this fruit dataset fit for the proposed model. However, the
precision of Apple Golden 1, Apple Red 3, and Apple Red Yellow 2 is 0.97, 0.96, and 0.98,
respectively. This is due to the fact that Apple Golden 1 resembles Apple Golden 2 in terms
of color; likewise, Apple Red 3 resembles Apple Red 2, and Apple Red Yellow 2 resembles
Apple Red Yellow 1. The overall accuracy of the proposed model for all classes of fruits
during training is 100%, as shown in the figure. It is important to note that applying
different preprocessing techniques in the model helped to achieve a high accuracy rate.
Usually, models perform well during training and under supervised data. However,
when it comes to real-world data, they usually do not perform well. For that reason, we
tested our proposed model’s performance by feeding it unseen data during the testing
phase. The dataset contains instances of images for testing purposes. It is important to
note that the model has not seen testing dataset images before. So, using such a dataset
for testing will help to identify the fair performance of the proposed model without any
bias. Table 4 presents the testing results of the TL-MobileNetV2 model in terms of precision,
recall, F1-score, and support. From the table, it can be inferred that the proposed model has
achieved the maximum accuracy in all classes except for a few, such as Apple Golden 1,
Apple Red 3, Apple Red Yellow 2, Banana Lady Finger, Cantaloupe 2, Cherry Wax Red,
and Kaki. Their precision is recorded 0.92, 0.96, 0.98, 0.92, 0.95, 0.92, and 0.94, respectively,
whereas the F1-score of these fruits is 0.94, 0.97, 0.96, 0.94, 0.96, 0.93, and 0.95, respectively.
This may be due to the fact that there are chances of false negative predictions between
Golden 1 and Golden 2, as well as between Red 1 and Red 3, and Red Yellow 1 and Red
Yellow 2 as these classes are somehow similar in terms of color. The model achieved
99% accuracy in both the training and testing phase. This also proves that the model did
not overfit as there would have been a difference in the training and testing score of the
TL-MobileNetV2 model.

Comparison of Model Performance

The proposed model was further compared with well-known models, as mentioned in
Section 2.2 (AlexNet [28], VGG16 [29], InceptionV3 [30], and ResNet [31]). To obtain a fair
comparison, the abovementioned model tuning and preprocessing techniques (discussed
in Section 2.4) were applied to the mentioned models with the same number of iterations.
Sustainability 2023, 15, 1906 10 of 14

Table 5 represents the performance of different models compared with the TL-MobileNetV2
model. It can be inferred that VGG16 performs worse in terms of accuracy, whereas
InceptionV3 performs slightly better than AlexNet and ResNet. It can be also noticed
that the MobileNetV2 is better than all of the other models, but the modified version of
MobileNetV2, TL-MobileNetV2, outperforms all the models mentioned in the table. TL-
MobileNetV2 is performing better in all the parameters whether it is precision, recall, or
F1-score. This is due to the fact that TL-MobileNetV2 is a modified version of MobileNetV2
with five extra layers, which played a great role in terms of improving the accuracy
of the model.

Table 3. Validation results of TL-MobileNetV2 model.

Fruit Types Precision Recall F1-Score Support

Apple Braeburn 1 1 1 4920
Apple Crimson Snow 1 1 1 4440
Apple Golden 1 0.97 1 0.98 4800
Apple Golden 2 1 1 1 4920
Apple Golden 3 1 1 1 4810
Apple Granny Smith 1 1 1 4920
Apple Pink Lady 1 1 1 4560
Apple Red 1 1 1 1 4920
Apple Red 2 1 1 1 4920
Apple Red 3 0.96 1 0.97 4290
Apple Red Delicious 1 1 1 4900
Apple Red Yellow 1 1 1 1 4920
Apple Red Yellow 2 0.98 1 0.99 6720
Apricot 1 1 1 4920
Avocado 1 1 1 4270
Banana 1 1 1 4900
Banana Lady Finger 0.98 1 0.99 4500
Banana Red 1 1 1 4900
Beetroot 1 1 1 4500
Blueberry 1 1 1 4620
Cactus Fruit 1 1 1 4900
Cantaloupe 1 1 1 1 4920
Cantaloupe 2 0.99 1 1 4920
Carambola 1 1 1 4920
Cherry 1 1 1 4920
Cherry Wax Black 1 1 1 4920
Cherry Wax Red 0.98 1 0.99 4920
Cherry Wax Yellow 1 1 1 4920
Chestnut 1 1 1 4500
Clementine 1 1 1 4900
Cocos 1 1 1 4900
Grape Blue 1 1 1 4900
Grape White 1 1 1 4920
Grapefruit Pink 1 1 1 4900
Grapefruit White 1 1 1 4920
Guava 1 1 1 4900
Huckleberry 1 1 1 4900
Kaki 0.97 1 0.99 4900
Kiwi 1 1 1 4660
Kumquats 1 1 1 4710
Accuracy - - 1 194,000
Macro Avg 1 1 1 194,000
Weighted Avg 1 1 1 194,000
Sustainability 2023, 15, 1906 11 of 14

Table 4. Testing results of TL-MobileNetV2 model.

Fruit Types Precision Recall F1-Score Support

Apple Braeburn 1 1 1 148
Apple Crimson Snow 1 1 1 160
Apple Golden 1 0.92 1 0.94 164
Apple Golden 2 1 1 1 161
Apple Golden 3 1 0.96 1 164
Apple Granny Smith 1 1 1 154
Apple Pink Lady 1 1 1 164
Apple Red 1 1 1 1 164
Apple Red 2 1 1 1 144
Apple Red 3 0.96 1 0.97 166
Apple Red Delicious 1 1 1 164
Apple Red Yellow 1 1 1 1 219
Apple Red Yellow 2 0.98 0.94 0.96 164
Apricot 1 1 1 143
Avocado 1 1 1 166
Banana 1 1 1 152
Banana Lady Finger 0.92 1 0.94 166
Banana Red 1 1 1 150
Beetroot 1 1 1 154
Blueberry 1 0.97 1 164
Cactus Fruit 1 1 1 164
Cantaloupe 1 1 1 1 166
Cantaloupe 2 0.95 1 0.96 164
Carambola 1 1 1 246
Cherry 1 1 1 164
Cherry Wax Black 1 1 1 164
Cherry Wax Red 0.92 1 0.93 153
Cherry Wax Yellow 1 1 1 166
Chestnut 1 1 1 166
Clementine 1 1 1 328
Cocos 1 1 1 164
Grape Blue 1 1 1 166
Grape White 1 1 1 166
Grapefruit Pink 1 0.99 1 166
Grapefruit White 1 1 1 166
Guava 1 1 1 166
Huckleberry 1 1 1 166
Kaki 0.94 1 0.95 164
Kiwi 1 1 1 156
Kumquats 1 1 1 157
Accuracy - - 0.99 6749
Macro Avg 0.99 0.99 0.99 6749
Weighted Avg 0.99 0.99 0.99 6749

Table 5. Performance of different models compared with the TL-MobileNetV2 model.

Models Precision Recall F1-Score

AlexNet 0.91 0.89 0.92
VGG16 0.88 0.90 0.88
InceptionV3 0.93 0.88 0.91
ResNet 0.89 0.92 0.90
MobileNetV2 0.96 0.97 0.96
TL-MobileNetV2 0.99 0.99 0.99

Furthermore, the proposed model, TL-MobileNetV2, was compared with start-of-the-

art models for its qualitative and quantitative features, as mentioned in Table 6. It can
Sustainability 2023, 15, 1906 12 of 14

be inferred from the table that different models were proposed for different datasets, for
instance, date fruit, pineapple, tomato, strawberry, and even mixed fruit datasets. Some
of these models adopted well-known architectures, such as ResNet or MobileNetV2, and
some created their models based on CNN. From the table, it can be concluded that our
proposed model performed well on the fruit dataset. The modified proposed model based
on MobileNetV2 has performed efficiently well when compared with other models and
is much faster and takes less space while processing. Additionally, this architecture is
best-suited for the collected dataset as it was collected using a mobile camera.

Table 6. Comparison of the TL-MobileNetV2 model with state-of-the-art approaches.

Paper Dataset Classes Method/Model Classification Type Accuracy (%)

[36] Mix Fruit 18 CNN Identification 95.08
[37] Date Fruit 8 MobileNetV2 Identification 99.0
[38] Pineapple 1 CNN Identification/quality 97.0
[39] Strawberry 1 ResNet Quality 84
[40] Tomato 1 SDF-convnets Quality 96.5
TL-MobileNetV2 Mix Fruit 40 MobileNetV2 Identification 99.0

4. Conclusions
Machine learning techniques, particularly those that are suited for computer vision,
have started to be widely used in precision agriculture. These techniques are used in various
areas, such as fruit classification, quality analysis, yield estimation, and disease prediction.
The success of these techniques has encouraged the development of deep learning models
for seed classification. In this study, a deep learning model TL-MobileNetV2 was developed
based on MobileNetV2 architecture. A dataset of forty types of fruits were used to train
and test the proposed model. In the TL-MobileNetV2 model, five different layers were
added after removing the classification layer present in the MobileNetV2 architecture to
improve the efficiency and accuracy of the model. Along with this, different preprocessing
and model-tuning techniques were used to make the TL-MobileNetV2 perform well on the
said dataset without overfitting. The experimental results show that the TL-MobileNetV2
has performed well on the fruit dataset by attaining 99% accuracy.
In future work, a mobile-based application will be further enhanced using a larger
number of different fruits, which aims to lead to a wider range of fruit classification. This
application will help people with limited knowledge to classify different types of fruit
and their different varieties. Furthermore, different CNN models will be trained on the
dataset, and their results will be compared to identify the best-fit model in terms of accuracy
and efficiency.

Funding: This research study was funded by the Deputyship for Research and Innovation, Ministry
of Education in Saudi Arabia, under project number INST199.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The dataset used in this study is public [27].
Acknowledgments: The author extends his appreciation to the Deputyship for Research and In-
novation, Ministry of Education in Saudi Arabia for funding this research through the project
number INST199.
Conflicts of Interest: The author declares that there is no conflict of interest regarding the publication
of this paper.
Sustainability 2023, 15, 1906 13 of 14

References
1. Gulzar, Y.; Khan, S.A. Skin Lesion Segmentation Based on Vision Transformers and Convolutional Neural Networks—A
Comparative Study. Appl. Sci. 2022, 12, 5990. [CrossRef]
2. Khan, S.A.; Gulzar, Y.; Turaev, S.; Peng, Y.S. A Modified HSIFT Descriptor for Medical Image Classification of Anatomy Objects.
Symmetry 2021, 13, 1987. [CrossRef]
3. Alam, S.; Raja, P.; Gulzar, Y. Investigation of Machine Learning Methods for Early Prediction of Neurodevelopmental Disorders
in Children. Wirel. Commun. Mob. Comput. 2022, 2022, 5766386. [CrossRef]
4. Sahlan, F.; Hamidi, F.; Misrat, M.Z.; Adli, M.H.; Wani, S.; Gulzar, Y. Prediction of Mental Health Among University Students. Int.
J. Perceptive Cogn. Comput. 2021, 7, 85–91.
5. Hanafi, M.F.F.M.; Nasir, M.S.F.M.; Wani, S.; Abdulghafor, R.A.A.; Gulzar, Y.; Hamid, Y. A Real Time Deep Learning Based Driver
Monitoring System. Int. J. Perceptive Cogn. Comput. 2021, 7, 79–84.
6. Kakani, V.; Nguyen, V.H.; Kumar, B.P.; Kim, H.; Pasupuleti, V.R. A Critical Review on Computer Vision and Artificial Intelligence
in Food Industry. J. Agric Food Res. 2020, 2, 100033. [CrossRef]
7. Behera, S.K.; Rath, A.K.; Mahapatra, A.; Sethy, P.K. Identification, Classification & Grading of Fruits Using Machine Learning &
Computer Intelligence: A Review. J. Ambient. Intell. Humaniz. Comput. 2020, 1–11. [CrossRef]
8. Altaheri, H.; Alsulaiman, M.; Muhammad, G. Date Fruit Classification for Robotic Harvesting in a Natural Environment Using
Deep Learning. IEEE Access 2019, 7, 117115–117133. [CrossRef]
9. Shamim Hossain, M.; Al-Hammadi, M.; Muhammad, G. Automatic Fruit Classification Using Deep Learning for Industrial
Applications. IEEE Trans. Ind. Inform. 2019, 15, 1027–1034. [CrossRef]
10. Gulzar, Y.; Hamid, Y.; Soomro, A.B.; Alwan, A.A.; Journaux, L. A Convolution Neural Network-Based Seed Classification System.
Symmetry 2020, 12, 2018. [CrossRef]
11. Hamid, Y.; Wani, S.; Soomro, A.B.; Alwan, A.A.; Gulzar, Y. Smart Seed Classification System Based on MobileNetV2 Architecture.
In Proceedings of the 2022 2nd International Conference on Computing and Information Technology (ICCIT), Tabuk, Saudi
Arabia, 25–27 January 2022; IEEE: Piscataway, NJ, USA, 2022; pp. 217–222.
12. Hamid, Y.; Elyassami, S.; Gulzar, Y.; Balasaraswathi, V.R.; Habuza, T.; Wani, S. An Improvised CNN Model for Fake Image
Detection. Int. J. Inf. Technol. 2022, 2022, 1–11. [CrossRef]
13. Saranya, N.; Srinivasan, K.; Pravin Kumar, S.K.; Rukkumani, V.; Ramya, R. Fruit Classification Using Traditional Machine Learning
and Deep Learning Approach; Springer: Cham, Switzerland, 2020; Volume 1108, ISBN 9783030372170.
14. Rojas-Aranda, J.L.; Nunez-Varela, J.I.; Cuevas-Tello, J.C.; Rangel-Ramirez, G. Fruit Classification for Retail Stores Using Deep
Learning; Springer: Cham, Switzerland, 2020; Volume 12088, ISBN 9783030490751.
15. Sridhar, B.; Kiran, K.S.S.; Likhitha, N.; Vardhan, K.P.V.A.; Nikhil, K.V. Development of an Automatic Fruit Classification Using Hybrid
Deep Learning Model for Super Markets; Springer: Singapore, 2022; Volume 838, ISBN 9789811685491.
16. Zhou, C.; Hu, J.; Xu, Z.; Yue, J.; Ye, H.; Yang, G. A Novel Greenhouse-Based System for the Detection and Plumpness Assessment
of Strawberry Using an Improved Deep Learning Technique. Front. Plant Sci. 2020, 11, 559. [CrossRef] [PubMed]
17. Mamat, N.; Othman, M.F.; Abdulghafor, R.; Alwan, A.A.; Gulzar, Y.; Malaysia, U.T.; Sultan, J.; Petra, Y. Enhancing Image
Annotation Technique of Fruit Classification Using a Deep Learning Approach. Sustainability 2023, 15, 901. [CrossRef]
18. Nasir, I.M.; Bibi, A.; Shah, J.H.; Khan, M.A.; Sharif, M.; Iqbal, K.; Nam, Y.; Kadry, S. Deep Learning-Based Classification of Fruit
Diseases: An Application for Precision Agriculture. Comput. Mater. Contin. 2020, 66, 1949–1962. [CrossRef]
19. Assuncao, E.; Diniz, C.; Gaspar, P.D.; Proenca, H. Decision-Making Support System for Fruit Diseases Classification Using Deep
Learning. In Proceedings of the 2020 International Conference on Decision Aid Sciences and Application, Sakheer, Bahrain, 8–9
November 2020; pp. 652–656.
20. Garillos-Manliguez, C.A.; Chiang, J.Y. Multimodal Deep Learning via Late Fusion for Non-Destructive Papaya Fruit Maturity
Classification. In Proceedings of the CCE 2021–2021 18th International Conference on Electrical Engineering, Computing Science
and Automatic Control, Mexico City, Mexico, 10–12 November 2021.
21. Herman, H.; Cenggoro, T.W.; Susanto, A.; Pardamean, B. Deep Learning for Oil Palm Fruit Ripeness Classification with Densenet.
In Proceedings of the 2021 International Conference on Information Management and Technology, Jakarta, Indonesia, 19–20
August 2021; pp. 116–119.
22. Suharjito; Elwirehardja, G.N.; Prayoga, J.S. Oil Palm Fresh Fruit Bunch Ripeness Classification on Mobile Devices Using Deep
Learning Approaches. Comput. Electron. Agric. 2021, 188, 106359. [CrossRef]
23. Mahmood, A.; Singh, S.K.; Tiwari, A.K. Pre-Trained Deep Learning-Based Classification of Jujube Fruits According to Their
Maturity Level. Neural. Comput. Appl. 2022, 34, 13925–13935. [CrossRef]
24. Musacchi, S.; Serra, S. Apple Fruit Quality: Overview on Pre-Harvest Factors. Sci. Hortic. 2018, 234, 409–430. [CrossRef]
25. SelectHealth.Org. Available online: https://selecthealth.org/blog/2020/02/how-many-types-of-apples-are-there-and-which-
is-best (accessed on 1 September 2022).
26. Sandler, M.; Howard, A.; Zhu, M.; Zhmoginov, A.; Chen, L.-C. Mobilenetv2: Inverted Residuals and Linear Bottlenecks. In
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–22 June 2018; pp.
4510–4520.
27. Oltean, M. Fruits 360 Dataset: A Dataset of Images Containing Fruits and Vegetables 2019. Kaggle, 2020. [Online]. Available
online: https://www.kaggle.com/moltean/fruits (accessed on 1 April 2022).
Sustainability 2023, 15, 1906 14 of 14

28. Krizhevsky, A.; Sutskever, I.; Hinton, G.E. Imagenet Classification with Deep Convolutional Neural Networks. Adv. Neural Inf.
Process. Syst. 2012, 25, 84–90. [CrossRef]
29. Simonyan, K.; Zisserman, A. Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv 2014, arXiv:1409.1556.
30. Szegedy, C.; Vanhoucke, V.; Ioffe, S.; Shlens, J.; Wojna, Z. Rethinking the Inception Architecture for Computer Vision. In
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 27–30 June 2016; pp.
2818–2826.
31. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep Residual Learning for Image Recognition. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778.
32. Ayoub, S.; Gulzar, Y.; Reegu, F.A.; Turaev, S. Generating Image Captions Using Bahdanau Attention Mechanism and Transfer
Learning. Symmetry 2022, 14, 2681. [CrossRef]
33. Zhuang, F.; Qi, Z.; Duan, K.; Xi, D.; Zhu, Y.; Zhu, H.; Xiong, H.; He, Q. A Comprehensive Survey on Transfer Learning. Proc. IEEE
2020, 109, 43–76. [CrossRef]
34. Tan, C.; Sun, F.; Kong, T.; Zhang, W.; Yang, C.; Liu, C. A Survey on Deep Transfer Learning. In Proceedings of the International
Conference on Artificial Neural Networks, Rhodes, Greece, 4–7 October 2018; Springer: Cham, Switzerland, 2018; pp. 270–279.
35. Arnold, T.B. KerasR: R Interface to the Keras Deep Learning Library. J. Open Source Softw. 2017, 2, 296. [CrossRef]
36. Zhang, Y.; Satapathy, S.C.; Wang, S. Fruit Category Classification by Fractional Fourier Entropy with Rotation Angle Vector Grid
and Stacked Sparse Autoencoder. Expert Syst. 2022, 39, e12701. [CrossRef]
37. Albarrak, K.; Gulzar, Y.; Hamid, Y.; Mehmood, A.; Soomro, A.B. A Deep Learning-Based Model for Date Fruit Classification.
Sustainability 2022, 14, 6339. [CrossRef]
38. Huang, T.-W.; Bhat, S.A.; Huang, N.-F.; Chang, C.-Y.; Chan, P.-C.; Elepano, A.R. Artificial Intelligence-Based Real-Time Pineapple
Quality Classification Using Acoustic Spectroscopy. Agriculture 2022, 12, 129. [CrossRef]
39. Su, Z.; Zhang, C.; Yan, T.; Zhu, J.; Zeng, Y.; Lu, X.; Gao, P.; Feng, L.; He, L.; Fan, L. Application of Hyperspectral Imaging for
Maturity and Soluble Solids Content Determination of Strawberry with Deep Learning Approaches. Front. Plant Sci. 2021, 12,
1897. [CrossRef]
40. Ko, K.; Jang, I.; Choi, J.H.; Lim, J.H.; Lee, D.U. Stochastic Decision Fusion of Convolutional Neural Networks for Tomato Ripeness
Detection in Agricultural Sorting Systems. Sensors 2021, 21, 917. [CrossRef] [PubMed]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.