Skip to main content
Loading...
Scholar9 logo True scholar network
  • Login/Sign up
  • Scholar9
    Publications ▼
    Article List Deposit Article
    Mentorship ▼
    Overview Sessions
    Q&A Institutions Scholars Journals
    Publications ▼
    Article List Deposit Article
    Mentorship ▼
    Overview Sessions
    Q&A Institutions Scholars Journals
  • Login/Sign up
  • Back to Top

    Transparent Peer Review By Scholar9

    OBJECT DETECTION USING YOLO

    Abstract

    Hey there! Let's talk about how important object detection is for self-driving cars. It helps them navigate safely and efficiently by recognizing and reacting to what's around them in real-time. This study looks into using the YOLO (You Only Look Once) algorithm, specifically YOLOv4, for spotting objects quickly in vehicles. YOLOv4 was chosen because it strikes a great balance between speed and accuracy, crucial for making fast decisions while driving autonomously. The model was trained on the COCO dataset, which has lots of different object classes, and then tested the KITTI dataset with realistic driving scenarios. They used metrics like precision, recall, F1-score, and mean Average Precision (mAP) to check how well YOLOv4 performed. Turns out it did a great job at detecting pedestrians, vehicles, and traffic signs even in tough situations like low light or when things are partially hidden. They also talked about how they got the data ready, trained the model, and made sure it can work in real-time. But even though YOLOv4 did really well, there are still some areas where it struggles, like finding small or hidden objects and handling all that math needed for quick processing. In a nutshell, this research shows that YOLOv4 can boost safety in self-driving cars by spotting objects effectively. It also points out ways to make detection algorithms better for the future, such as improving the model design and combining data from different sensors.

    Reviewer Photo

    Shyamakrishna Siddharth Chamarthy Reviewer

    badge Review Request Accepted
    Reviewer Photo

    Shyamakrishna Siddharth Chamarthy Reviewer

    10 Oct 2024 06:33 PM

    badge Approved

    Relevance and Originality

    Methodology

    Validity & Reliability

    Clarity and Structure

    Results and Analysis

    Relevance and Originality

    The study is highly relevant in the context of the growing interest in autonomous driving technologies. Object detection plays a crucial role in ensuring the safety and efficiency of self-driving cars by enabling them to recognize and respond to their environment in real-time. The focus on the YOLO (You Only Look Once) algorithm, particularly YOLOv4, is original as it highlights an innovative approach that balances speed and accuracy. By demonstrating YOLOv4's capabilities in detecting various objects under challenging conditions, the research contributes valuable insights into enhancing the reliability of autonomous vehicles.


    Methodology

    The methodology employed in the study is robust, leveraging established datasets such as COCO for training and KITTI for testing. This approach is effective in evaluating the model's performance in realistic driving scenarios. The use of precision, recall, F1-score, and mean Average Precision (mAP) as evaluation metrics is appropriate, as these metrics provide a comprehensive understanding of the model's effectiveness. However, further details on the preprocessing steps and data augmentation techniques used to enhance the dataset would strengthen the methodology section and provide a clearer picture of how the model was prepared for training.


    Validity & Reliability

    The validity of the findings is supported by the successful performance of YOLOv4 in detecting pedestrians, vehicles, and traffic signs, even in difficult conditions like low light and partial occlusions. These results suggest that the model is reliable for real-world applications in self-driving cars. However, to enhance reliability, the study should include a discussion on the model's performance across different environmental conditions and scenarios. Additionally, comparing YOLOv4 with other object detection algorithms in similar conditions would further validate its effectiveness.


    Clarity and Structure

    The article is well-structured, clearly articulating the significance of object detection and the advantages of using YOLOv4. The explanation of technical terms and processes is accessible, making it easy for readers to follow the research's implications. Nevertheless, incorporating visual elements such as flowcharts or graphs depicting the model's performance metrics would enhance clarity. A summary section at the end of the article could also help consolidate the main findings and their implications for future research.


    Result Analysis

    The result analysis is effective in demonstrating YOLOv4's strengths in object detection for self-driving cars. The mention of specific detection capabilities, even in challenging environments, showcases the model's practicality. However, the analysis could be deepened by discussing the model's limitations in detecting small or hidden objects and the computational challenges associated with real-time processing. Addressing these areas could provide a more balanced view of the model's performance and inform future enhancements. Overall, the research highlights the potential of YOLOv4 in improving safety and efficiency in autonomous driving, suggesting directions for further advancements in detection algorithms and sensor integration.

    Publisher Logo

    IJ Publication Publisher

    Thank you sir

    Publisher

    IJ Publication

    IJ Publication

    Reviewer

    Shyamakrishna Siddharth

    Shyamakrishna Siddharth Chamarthy

    More Detail

    Category Icon

    Paper Category

    Computer Engineering

    Journal Icon

    Journal Name

    JETIR - Journal of Emerging Technologies and Innovative Research External Link

    Info Icon

    p-ISSN

    Info Icon

    e-ISSN

    2349-5162

    Subscribe us to get updated

    logo logo

    Scholar9 is aiming to empower the research community around the world with the help of technology & innovation. Scholar9 provides the required platform to Scholar for visibility & credibility.

    QUICKLINKS

    • What is Scholar9?
    • About Us
    • Mission Vision
    • Contact Us
    • Privacy Policy
    • Terms of Use
    • Blogs
    • FAQ

    CONTACT US

    • +91 82003 85143
    • hello@scholar9.com
    • www.scholar9.com

    © 2026 Sequence Research & Development Pvt Ltd. All Rights Reserved.

    whatsapp