DOI QR코드

DOI QR Code

A Scene-Specific Object Detection System Utilizing the Advantages of Fixed-Location Cameras

  • Jin Ho Lee (School of Computer Science and Engineering, Kyungpook National University) ;
  • In Su Kim (School of Computer Science and Engineering, Kyungpook National University) ;
  • Hector Acosta (School of Computer Science and Engineering, Kyungpook National University) ;
  • Hyeong Bok Kim (Testworks, Inc.) ;
  • Seung Won Lee (Testworks, Inc.) ;
  • Soon Ki Jung (School of Computer Science and Engineering, Kyungpook National University)
  • Received : 2023.04.27
  • Accepted : 2023.09.27
  • Published : 2023.12.31

Abstract

This paper introduces an edge AI-based scene-specific object detection system for long-term traffic management, focusing on analyzing congestion and movement via cameras. It aims to balance fast processing and accuracy in traffic flow data analysis using edge computing. We adapt the YOLOv5 model, with four heads, to a scene-specific model that utilizes the fixed camera's scene-specific properties. This model selectively detects objects based on scale by blocking nodes, ensuring only objects of certain sizes are identified. A decision module then selects the most suitable object detector for each scene, enhancing inference speed without significant accuracy loss, as demonstrated in our experiments.

Keywords

Acknowledgement

This research was supported by the MSIT (Ministry of Science and ICT), Korea, under the Innovative Human Resource Development for Local Intellectualization support program (IITP-2023-RS-2022-00156389) supervised by the IITP (Institute for Information & Communications Technology Planning & Evaluation) and also was supported by the BK21 FOUR project (AI-driven Convergence Software Education Research Program) funded by the Ministry of Education, School of Computer Science and Engineering, Kyungpook National University, Korea (4199990214394).

References

  1. D. B. Nguyen, C. R. Dow, and S. F. Hwang, "An efficient traffic congestion monitoring system on internet of vehicles," Wireless Communications and Mobile Computing, vol. 2018, pp. 1-17, 2018. DOI: 10.1155/2018/9136813.
  2. G. P. Rocha Filho, R. I. Meneguette, J. R. Torres Neto, A. Valejo, L. Weigang, J. Ueyama, G. Pessin, and L. A. Villas, "Enhancing intelligence in traffic management systems to aid in vehicle traffic congestion problems in smart cities," Ad Hoc Networks, vol. 107, p. 102265, Oct. 2020. DOI: 10.1016/j.adhoc.2020.102265.
  3. K. H. N. Bui, H. Yi, and J. Cho, "A multi-class multi-movement vehicle counting framework for traffic analysis in complex areas using CCTV systems," Energies, vol. 13, no. 8, p. 2036, Apr. 2020. DOI: 10.3390/en13082036.
  4. M. Sadiq, S. Masood, and O. Pal, "FD-YOLOv5: A fuzzy image enhancement based robust object detection model for safety helmet detection," International Journal of Fuzzy Systems, vol. 24, no. 5, pp. 2600-2616, Jul. 2022. DOI: 10.1007/s40815-022-01267-2.
  5. G. W. Chen, Y. H. Lin, M. T. Sun, and T. U. Ik, "Managing edge AI cameras for traffic monitoring," in 2022 23rd Asia-Pacific Network Operations and Management Symposium (APNOMS), Takamatsu, Japan, pp. 01-04, 2022. DOI: 10.23919/APNOMS56106.2022.9919965.
  6. K. Y. Cao, Y. F. Liu, G. J. Meng, and Q. M. Sun, "An overview on edge computing research," IEEE Access, vol. 8, pp. 85714-85728, 2020. DOI: 10.1109/Access.2020.2991734.
  7. M. Gusev and S. Dustdar, "Going back to the roots-the evolution of edge computing, an IoT perspective," Ieee Internet Computing, vol. 22, no. 2, pp. 5-15, Mar. 2018. DOI: 10.1109/Mic.2018.022021657.
  8. X. F. Wang, Y. W. Han, V. C. M. Leung, D. Niyato, X. Q. Yan, and X. Chen, "Convergence of edge computing and deep learning: A comprehensive survey," IEEE Communications Surveys & Tutorials, vol. 22, no. 2, pp. 869-904, 2020. DOI: 10.1109/Comst.2020.2970550.
  9. A. Mhalla, T. Chateau, S. Gazzah, and N. E. Ben Amara, "An embedded computer-vision system for multi-object detection in traffic surveillance," IEEE Transactions on Intelligent Transportation Systems, vol. 20, no. 11, pp. 4006-4018, Nov. 2019. DOI: 10.1109/Tits.2018.2876614.
  10. Y. Jeong, H. W. Oh, S. Kim, and S. E. Lee, "An edge AI device based intelligent transportation system," Journal of information and communication convergence engineering, vol. 20, no. 3, pp. 166-173, Sep. 2022. DOI: 10.56977/jicce.2022.20.3.166.
  11. S. Q. Ren, K. M. He, R. Girshick, and J. Sun, "Faster R-CNN: Towards real-time object detection with region proposal networks," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 6, pp. 1137-1149, Jun. 2017. DOI: 10.1109/Tpami.2016.2577031.
  12. K. M. He, G. Gkioxari, P. Dollar, and R. Girshick, "Mask R-CNN," in Proceedings of the IEEE international conference on computer vision, Venice, Italy, pp. 2980-2988, 2017. DOI: 10.1109/Iccv.2017.322.
  13. J. Redmon and A. Farhadi, "YOLO9000: Better, faster, stronger," in 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, USA, pp. 6517-6525, 2017. DOI: 10.1109/Cvpr.2017.690.
  14. J. Redmon and A. Farhadi, "Yolov3: An incremental improvement," arXiv preprint arXiv:1804.02767, Apr. 2018. DOI: 10.48550/arXiv.1804.02767.
  15. A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, "Yolov4: Optimal speed and accuracy of object detection," arXiv preprint arXiv: 2004.10934, Apr. 2020. DOI: 10.48550/arXiv.2004.10934.
  16. G. JOCHER, YOLOv5, ed, 2020. [Online], Available: https://github.com/ultralytics/yolov5.
  17. C. Y. Wang, H. Y. M. Liao, Y. H. Wu, P. Y. Chen, J. W. Hsieh, and I. H. Yeh, "CSPNet: A new backbone that can enhance learning capability of CNN," in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, USA, pp. 1571-1580, 2020. DOI: 10.1109/Cvprw50498.2020.00203.
  18. S. Liu, L. Qi, H. F. Qin, J. P. Shi, and J. Y. Jia, "Path aggregation network for instance segmentation," in 2018 Ieee/Cvf Conference on Computer Vision and Pattern Recognition (Cvpr), Salt Lake City, USA, pp. 8759-8768, 2018. DOI: 10.1109/Cvpr.2018.00913.
  19. J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You only look once: Unified, real-time object detection," in 2016 Ieee Conference on Computer Vision and Pattern Recognition (Cvpr), Las Vegas, USA, pp. 779-788, 2016. DOI: 10.1109/Cvpr.2016.91.
  20. L. Wang and K. J. Yoon, "Knowledge distillation and student-teacher learning for visual intelligence: A review and new outlooks," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 6, pp. 3048-3068, Jun. 2022. DOI: 10.1109/Tpami.2021.3055564.
  21. H. Vanholder, "Efficient inference with tensorrt," in GPU Technology Conference, vol. 1, p. 2. 2016.
  22. Y. Gong, L. Liu, M. Yang, and L. Bourdev, "Compressing deep convolutional networks using vector quantization," arXiv preprint arXiv:1412.6115, Dec. 2014. DOI: 10.48550/arXiv.1412.6115.
  23. S. Han, J. Pool, J. Tran, and W. Dally, "Learning both weights and connections for efficient neural network," Advances in neural information processing systems, vol. 28, 2015.
  24. X. K. Zhu, S. C. Lyu, X. Wang, and Q. Zhao, "TPH-YOLOv5: Improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios," in 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), Montreal, Canada, pp. 2778-2788, 2021. DOI: 10.1109/Iccvw54120.2021.00312.
  25. Y. Li, X. Y. Bai, and C. L. Xia, "An improved YOLOV5 based on triplet attention and prediction head optimization for marine organism detection on underwater mobile platforms," Journal of Marine Science and Engineering, vol. 10, no. 9, p. 1230, Sep. 2022. DOI: 10.3390/jmse10091230.