https://doi.org/10.46335/IJIES.2024.9.1.

20 e-ISSN: 2456-3463
Vol. 9, No. 1, 2024, PP. 108-112

International Journal of Innovations in Engineering and Science, www.ijies.net

Object Detection Using AI

Neha Shinde1, Sandali Tagunde2, Saraswati Zerkunte3, Prof. Megha Dhotay4

1
Dept. Computer engineering, MIT Polytechnic Kothrud-411038, Pune, India
MIT Polytechnic, Pune, Kothrud 411038

*Corresponding Authors: [email protected], [email protected]

Received on: 17 March, 2024 Revised on: 20 April, 2024 Published on: 22 April, 2024

Abstract- Object detection using artificial intelligence systematic approach to tackle this complex task. The
(AI) has revolutionized various industries, from
process begins with a clear definition of the object
surveillance to autonomous vehicles and retail. This
paper provides a concise overview of object detection detection problem, collection and annotation follow,
techniques employing AI algorithms. It explores the where a diverse dataset of gathered and labelled. Data
fundamental concepts behind object detection, including preprocessing steps valve images or videos containing
the use of convolutional neural networks (CNNs) and the target objects is resizing, normalization.
deep learning architectures. The abstract discusses key
methodologies such as region based convolutional
Camera Based Detection:-
neural networks (RCNN), single shot detectors (SSD),
and You Only Look Once (YOLO). Additionally, it Camera-based object detection using AI involves using
highlights recent advancements, challenges, and future a machine learning model to identify objects within
directions in Aid riven object detection. Understanding images or video frames captured by a camera in real-
these approaches is critical for researchers and time. The resources been used for developing this mouse
practitioners seeking to harness the power of AI for system are these:
robust and efficient object detection systems.
• HTML CSS JavaScript Software:-
Keywords: Object localization, Object recognition,
Image Preprocessing, Leaf disease detection.
HTML (Hypertext Markup Language) is the standard
markup language used to create and design web pages.
I. INTRODUCTION CSS (Cascading Style Sheets) is a style sheet language
used to describe the presentation of HTML elements on
Object detection using AI is a cutting-edge technology a. web page.
JavaScript is a programming language commonly used to
that empowers machines to recognize and locate objects create interactive and dynamic elements on web page.
within images or video streams. It plays a pivotal role in
computer vision, allowing computers to understand the
visual world much like humans do. At the heart of this
technology are Convolutional Neural Networks (CNNs),
which have proven exceptionally effective in image
analysis. These networks are trained to identify specific
objects within an image and then draw bounding boxes.
Around them, providing both their position and class
labels, such as "car," "person," or "cat." Proposing a
methodology for object detection using AI involves a Fig 1. HTML CSS JS Software

108
https://doi.org/10.46335/IJIES.2024.9.1.20 e-ISSN: 2456-3463
Vol. 9, No. 1, 2024, PP. 108-112

International Journal of Innovations in Engineering and Science, www.ijies.net

Uses In Different Fields: models face difficulty in deployment on mobile

terminals. In this case, it is of high practical significance
1. Autonomous Driving and Vehicle Detection and application value to carry out the research on the
lightweight of the recognition model.
2. Surveillance and Security
3. Healthcare and Medical Imaging III. SYSTEM ARCHITECTURE
4. Object Recognition in Images and Videos
 One popular deep learning model for object
detection is the Convolutional Neural Network
(CNN).CNNs have been extensively used in
various object detection frameworks due to their
ability to automatically learn hierarchical features
from input images.
 It efficiently generates region proposals and then
classifies those proposals using a CNN. This
approach significantly speeds up the detection
process compared to earlier R-CNN models.
 These CNN-based models have been successfully
applied in various domains, including object
Fig .2.Surveillance and Security detection in cameras and agriculture, due to their
ability to detect objects accurately and efficiently.
II. LITERATURE SURVEY
The Basic block diagram of the system:
[1] In recent years, there has been a growing need to
perform object detection at the edge. Since the edge
environment has tight physical constraints, the efficient
use of AI devices is a key challenge to execute object
detection at high throughput.
[2] In this paper, we propose an object detection
acceleration method which uses two types of one-stage
detectors in combination. After detecting object
candidates by a lightweight detector, the method
generates aggregated images by combining the candidate
images and executes the second more accurate detector
on the aggregated images to improve execution
efficiency of AI devices.
[3] Our evaluations confirmed that the proposed method
can speed up object detection by up to eight times for a
license plate detection task with almost no accuracy
degradation. We conducted evaluations with a car
detection task and a pose estimation task as well and
confirmed the broad applicability of the proposed
method. For object detection in garbage sorting Fig.3.Block Diagram
applications, the YOLOv4 model have been employed in
the detection of abnormal objects during waste sorting. IV- PROPOSED METHODOLOGY
YOLO models have been proven as an effective solution
with stable performance and accuracy. Capturing the real time view:
[4] In addition, mobile terminals such as smart cameras
have limited computing resources, and conventional AI

109
https://doi.org/10.46335/IJIES.2024.9.1.20 e-ISSN: 2456-3463
Vol. 9, No. 1, 2024, PP. 108-112

International Journal of Innovations in Engineering and Science, www.ijies.net

a) Model Loading: Load a pre-trained object detection

model.
b) Initialization: Prepare the model for inference.
c) Input Acquisition: Capture input data (e.g., images or
video frames).
d) Preprocessing: Prepare the input data for the model
(e.g., resizing, normalization).
e) Inference: Analyse the input data to identify objects
present.
f) Detection Output: Generate predictions with
bounding boxes and labels.
g) Postprocessing: Refine the detection results if needed
(e.g., filtering, non-maximum suppression).
h) Visualization: Display the detected objects on the Fig.5.Output
input data.
i) Continuous Operation: Continuously process new
As the camera captures video frames, they are processed
input data for real-time detection.
by the object detection model, which identifies objects
j) User Interaction: Allow users to adjust parameters or
present in the scene. Detected objects are displayed on
enable/disable features for customization.
the video feed with bounding boxes and labels,
indicating their positions and classifications. Users can
interact with the system by adjusting parameters to
customize the detection behaviour. Additionally, the
system provides a reset button to clear any settings or
detected objects. Overall, the code creates an interactive
AI object detection interface, allowing users to observe
and analyse their surroundings in real-time using
computer vision technology.

Agriculture:

In recent years, AI has been widely adopted in

agriculture to improve productivity and efficiency.
Object detection is one such technique that has gained
Fig.4.Object Detection using Camera immense importance in the field of agriculture.
Object detection is a valuable tool in agriculture because
Working:
it allows for more efficient and precise farming practices,
In these project it implements an AI object detection
leading to increased productivity and sustainability while
system using a pre-trained model called "cocossd" and
reducing costs and environmental impact.
the ml5.js library. When the webpage loads, it displays a
user interface consisting of a video element for live
camera feed, input fields for custom object labels and
confidence threshold, and a toggle switch to
enable/disable AI object detection. The AI model is
loaded and initialized, allowing it to process input data
from the camera in real-time. Users can enable the AI
detection by toggling the switch and adjust parameters
like FPS (frames per second) and confidence threshold.

Fig.6.Agriculture

110
https://doi.org/10.46335/IJIES.2024.9.1.20 e-ISSN: 2456-3463
Vol. 9, No. 1, 2024, PP. 108-112

International Journal of Innovations in Engineering and Science, www.ijies.net

DISEASE DETECTION simulate Detection() function would be replaced with

genuine detection logic, potentially leveraging machine
The provided code forms the foundation of a web-based learning or image processing algorithms to accurately
application designed for agricultural purposes, identify plant diseases based on uploaded images.
specifically focusing on plant disease detection. The
HTML structure defines the user interface, featuring an
image upload form where users can submit images of
plants for analysis. Upon selecting an image and clicking
the "Detect Disease" button, the JavaScript function
detect Disease() is invoked. Currently, this function
initiates a placeholder simulation of disease detection,
presenting mock results within the designated result
container. The detected region seems like the following
image:

Fig.8. Leaf disease detection

APPLICATION

This technology has wide applications in the fields of

Autonomous Driving and Vehicle Detection,
Surveillance and Security, Healthcare and Medical
Imaging, Object Recognition in Images and Videos.

FUTURE SCOPE
Overall, the future holds promise for object detection AI
to revolutionize various sectors and enhance our daily
lives with intelligent and reliable detection capabilities.

CONCLUSION
Fig.7.Output The system demonstrates the potential of AI in practical
 It includes UI elements for controlling the detection applications, such as surveillance, inventory management,
process, such as toggling AI, adjusting FPS, and and augmented reality. As AI technologies continue to
entering custom object labels. evolve, this implementation serves as a foundation for
 JavaScript is used to handle model loading, user developing more sophisticated object detection solutions
interactions, camera feed processing, and object across various domains, contributing to advancements in
detection. computer vision.
 CSS is used for styling the layout and appearance
of HTML element. REFERENCES

PLANT DISEASE DATASET [1] Watanabe, Yoshikazu, et al. "Work-in-progress:

Object detection acceleration method by improving
The CSS styling enhances the visual presentation of the execution efficiency of ai device." 2022
webpage, including elements such as the background International Conference on Compilers,
image, navigation bar, and button. However, the Architecture, and Synthesis for Embedded Systems
functionality of the application remains limited to the (CASES). IEEE, 2022.
provided simulation, where the actual disease detection [2] Wei, Zhang, et al. "Efficient object detection of non-
process is not implemented. In a real-world scenario, the perishable in smart garbage sorting booths using

111
https://doi.org/10.46335/IJIES.2024.9.1.20 e-ISSN: 2456-3463
Vol. 9, No. 1, 2024, PP. 108-112

International Journal of Innovations in Engineering and Science, www.ijies.net

the Efficient-Net model." 2022 China Automation

Congress (CAC). IEEE, 2022.
[3] Bey-Temsamani, Abdellatif, Agusmian Partogi
Ompusunggu, and Florian Hammer. "AI meets
Agriculture: A Smart System for Foreign Object
Damage Avoidance." 2020 IEEE Conference on
Industrial Cyberphysical Systems (ICPS). Vol. 1.
IEEE, 2020.
[4] Rani, Ruchi, et al. "Role of Artificial Intelligence in
Agriculture: An Analysis and Advancements with
Focus on Plant Diseases." IEEE Access (2023).
[5] Guo, Chaoxu, et al. "Augfpn: Improving multi-
scale feature learning for object
detection." Proceedings of the IEEE/CVF
conference on computer vision and pattern
recognition. 2020.
[6] Ramachandran, Anitha, and Arun Kumar
Sangaiah. "A review on object detection in
unmanned aerial vehicle
surveillance." International Journal of Cognitive
Computing in Engineering 2 (2021): 215-228.

112