Unveiling the Potential of YOLOv9 through Comparison with YOLOv8
Keywords:
YOLOv9, YOLOv8, Object Detection, Training Efficiency, Accuracy.Abstract
This study introduces YOLOv9, a new object detection model building upon the success of YOLOv8. While YOLOv8 has delivered impressive results, there was a push to further enhance accuracy and efficiency. We delve into the architectures of both YOLO models to understand the trade-off between training speed and accuracy. YOLOv9 introduces additional complexity compared to its predecessor, which translates to superior performance. To evaluate the models' capabilities, we leveraged a tomato disease detection dataset from Roboflow. This dataset encompasses three disease classes for tomato fruits, along with a healthy class. Our experiments demonstrate that YOLOv9 achieves a significant improvement in accuracy (93.6% vs. 92%), while maintaining comparable training efficiency. However, it requires slightly longer training times compared to YOLOv8. To further substantiate these results, we present comprehensive analyses of precision, recall, F1-score, and loss functions during both training and validation stages. Additionally, in the testing phase, YOLOv9 exhibits superior precision in detecting tomato diseases. While requiring slightly more training resources, YOLOv9 offers a compelling trade-off between accuracy and efficiency. This makes it a promising choice for applications where precise object detection is paramount.
Downloads
References
J. Redmon & Ali Farhadi. “YOLOv3: An Incremental Improvement. Computer Vision and Pattern Recognition”, in Computer Vision and Pattern Recognition , 2018.
Lu Tan, Tianran Huangfu, Liyao Wu & Wenying Chen. “Comparison of RetinaNet, SSD, and YOLO v3 for real-time pill identification”, in BMC Medical Informatics and Decision Making, 2021, vol. 21, p. 324.
Alexey Bochkovskiy, Chien-Yao Wang, Hong-Yuan Mark Liao. “YOLOv4: Optimal Speed and Accuracy of Object Detection”, in Computer Vision and Pattern Recognition, 2020.
Muhammad Abdullah. “YOLO Working principle, difference between its different Variants and Versions”, in Medium, 2023, https://medium.com/ @muhabd51/ yolo-working-principle-difference-between-its-ddifferent-variants-and-versions-95b8ad7b95ab.
Zhongqiang Luo, Chenghao Wang, Ziyuan Qi, Chunlan Luo, “A_YOLOv8s: A lightweight-attention YOLOv8s for oil leakage detection in power transformers”, in Alexandria Engineering Journal, 2024, vol. 92, pp. 82--91.
Chien-Yao Wang, I-Hau Yeh, Hong-Yuan Mark Liao, “YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information”, in Computer Vision and Pattern Recognition, arXiv:2402.13616, 2024.
Haitong Lou, Xuehu Duan, Junmei Guo, Haiying Liu, Jason Gu, Lingyun Bi, Haonan Chen, “DC-YOLOv8: Small size Object detection algorithm based on camera sensor”, 2023, doi:10.20944/preprints202304.0124.v1
Moahaimen Talib, Ahmed H. Y. Al-Noori, Jameelah Suad, “YOLOv8-CAB: Improved YOLOv8 for Real-time object detection”, in Karbala International Journal of Modern Science, 2024.
Savan Agrawal, “Helmet Detection YOLOv3: A YOLOv3 detector, which can detect helmet”, in Kaggle, 2020.
Krunal Patel, Vrajesh Patel, Vikrant Prajapati, Darshak Chauhan, Adil Haji, Sheshang Degadwala, “Safety Helmet Detection Using YOLOV8” in International Conference on Pervasive Computing and Social Networking (ICPCSN), 2023, DOI: 10.1109/ICPCSN58827.2023. 00012.
S. Roy, S. Mukherjee, S. K. Ghosh, “YOLO Based Real-Time Object Detection for Video Surveillance”, in 3rd International Conference on Advanced Computing and Communication Systems (ICACCS), 2019, pp. 1537-1542.
G. R. Goswami, A. K. Singh, H. K. Bajaj, “Real-Time Object Detection for Security Applications Using YOLOv3”, in 4th International Conference on Signal Processing, Computing and Control, 2019, pp. 147-152, https://ieeexplore.ieee. org/document/8918322.
X. Zhou, Y. Yao, G. Liu, Z. Sun, Y. Hu, “Deep learning for fine-grained disease classification of tomato leaves”, in IEEE Access, 2019, 107386-107400, https://ieeexplore.ieee.org/ document/ 8804202.
Santosh Adhikari, Bikesh Shrestha, Bibek Baiju, Er. Saban Kumar K.C, “Tomato Plant Diseases Detection System using Image Processing”, in 1st Kantipur Engineering College, Dhapakhel, Lalitpur Conference Proceedings, 2018.
Md Ershadul Haque, Ashikur Rahman, Iftekhar Junaeid, Samiul Ul Hoque, Manoranjan Paul, “Rice Leaf Disease Classification and Detection using YOLOV5”, 2022, arXiv:2209.01579v.
Md. Janibul Alam Soeb, Md. Fahad Jubayer, Tahmina Akanjee Tarin, Muhammad Rashed Al Mamun, Fahim Mahafuz Ruhad, Aney Parven, Nabisab Mujawar Mubarak, Soni Lanka Karri Islam Md. Meftaul, “Tealeaf disease detection and identification based on YOLOv7 (YOLO‑T)”, in Scientific Reports, 2023, https://doi.org/10.1038/ s41598-023-33270-4.
Mubashiru Olarewaju Lawal, “Tomato detection based on modified YOLOv3 framework”, in Scientific Reports, 2021, https://doi.org/10. 1038/s41598-021-81216-5.
M. R. Shams, S. A. Sharief, M. H. Abdullah, “Performance Analysis of Deep Learning Techniques for Rice Disease Detection”, in IEEE Access, 2021, 13132-13141, https://ieeexplore. ieee.org/ document/9414761.
Tran Quang Vinh, Haewon Byeon, Enhancing “Alzheimer's Disease Diagnosis: The Efficacy of the YOLO Algorithm Model”, in International Journal of Advanced Computer Science and Applications(IJACSA), 2023, vol. 14, Issue 11, DOI: 10.14569/IJACSA.2023.0141182.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
