Enhanced YOLOv8 with BiFPN and MHSA for traffic vehicle detection

https://doi.org/10.31130/ud-jst.2025.23(9D).558E

Tóm tắt: 344 | PDF: 121

##plugins.themes.academic_pro.article.main##

Author

  • Ngo Truong An
    Master’s Student, Computer Science, The University of Danang - University of Science and Technology, Vietnam; Dong A University, Vietnam
    Huynh Huu Hung
    The University of Danang - University of Science and Technology, Vietnam
    Tran Thi Hoang Oanh
    Dong A University, Vietnam

Từ khóa:

YOLOv8 architecture
object detection
bidirectional feature pyramid network (BiFPN)
multi-head self-attention (MHSA)

Tóm tắt

Accurately detecting vehicles in urban traffic scenarios is a complex task, especially when dealing with cluttered backgrounds, diverse object scales, and high vehicle density. In this study, we propose an improved YOLOv8-based model tailored for vehicle detection in such challenging environments. The enhancement lies in the integration of a Bidirectional Feature Pyramid Network (BiFPN), which boosts multi-scale feature fusion, and a Multi-Head Self-Attention (MHSA) module, designed to strengthen the model’s capacity to understand broader spatial context. Together, these components help the model better distinguish between densely arranged vehicles. We conducted in-depth experiments on the Vehicles-COCO dataset, and the results demonstrate that our YOLOv8-BiFPN-MHSA variant outperforms the original YOLOv8 not only in Precision but also in mAP. Our model achieves significantly higher mAP@0.5 and mAP@0.5:0.95, along with an overall improvement in detection performance. These enhancements highlight the stability, efficiency, and strong potential for real-world traffic monitoring systems.

Tài liệu tham khảo

    [1] C. Chen et al., "Enhanced YOLOv5: An Efficient Road Object Detection Method," Sensors, vol. 23, no. 20, art. 8355, 2023.
    [2] J. Redmon et al., “You Only Look Once: Unified, Real-Time Object Detection”, in Proc. IEEE CVPR, 2016, pp. 779–788.
    [3] N. U. A. Tahir et al., “PVswin-YOLOv8s: UAV-Based Pedestrian and Vehicle Detection for Traffic Management”, Drones, vol. 8, no. 3, art. 84, 2024.
    [4] H. Guo et al., “Research on Vehicle Detection Based on Improved YOLOv8 Network”, arXiv:2501.00300, 2025.
    [5] X. Liu et al., “YOLOv8-FDD: A Real-Time Vehicle Detection Method Based on Improved YOLOv8”, IEEE Access, vol. 12, pp. 136280–136296, 2024.
    [6] M. Tan, R. Pang, and Q. V. Le, “Eff1icientDet: Scalable and Efficient Object Detection”, in Proc. IEEE CVPR, 2020, pp. 10781–10790.
    [7] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, "Attention is all you need," in Advances in Neural Information Processing Systems, vol. 30, pp. 5998–6008, 2017.
    [8] C. Lv et al., “Vehicle detection and classification using an ensemble of EfficientDet and YOLOv8”, PeerJ Comput. Sci., vol. 10, art. e2233, 2024.
    [9] B. Yang, J. Yan, Z. Lei, and S. Z. Li, “Convolutional channel features”, in Proc. IEEE Int. Conf. Comput. Vis., 2015, pp. 82–90.
    [10] G. Mandellos, I. Keramitsoglou, and C. Kiranoudis, “A background subtraction algorithm for detecting and tracking vehicles”, Expert Syst. Appl., vol. 38, no. 3, pp. 1619–1631, 2011.
    [11] J. Cao, W. Zhang, Z. He, and H. Sun, “Robust vehicle detection in aerial images based on convolutional neural networks”, IEEE Access, vol. 9, pp. 45620–45629, 2021.
    [12] M. T. Islam, R. S. Amin, and M. Murshed, “A comprehensive review of deep learning-based approaches for traffic object detection”, Sensors, vol. 24, no. 1, p. 222, 2024.
    [13] N. Li, S. Tang, Y. Fan, and Y. Zhang, “YOLOv8-MECA-ASPF: A lightweight vehicle detection algorithm based on multi-attention and context perception”, in Proc. 2024 IEEE Int. Conf. Intell. Transp. Syst. (ITSC), 2024, pp. 1–6.
    [14] H. Li, Y. Song, and X. Sun, “Lightweight network design for efficient object detection using MECA and ASPF modules”, Neurocomputing, vol. 540, pp. 25–34, 2023.
    [15] L. Wen, D. Du, Z. Cai, Z. Lei, M. Chang, H. Qi, J. Lim, M. Yang, and S. Lyu, “UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking”, Comput. Vis. Image Underst., vol. 193, pp. 102907, 2020.
    [16] A. Geiger, P. Lenz, and R. Urtasun, “Are we ready for autonomous driving? The KITTI vision benchmark suite”, in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2012, pp. 3354–3361.
    [17] Y. Zhu, C. Zhao, J. Wang, X. Zhao, Y. Wu, and H. Lu, “Vision meets drones: A challenge”, in Proc. Eur. Conf. Comput. Vis. (ECCV), 2018, pp. 643–659.
    [18] X. Wang et al., “Non-Local Neural Networks”, in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2018, pp. 7794–7803.
    [19] J. Terven and D. M. Cordova-Esparza, “A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS”, Machine Learning and Knowledge Extraction, vol. 5, no. 4, pp. 1680–1716, 2023.
    [20] I. Bello et al., “Attention Augmented Convolutional Networks”, in Proc. IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 3286–3295.
    [21] A. Dosovitskiy et al., “An Image is Worth 16×16 Words: Transformers for Image Recognition at Scale”, in Proc. Int. Conf. on Learning Representations (ICLR), 2021.
    [22] T.-Y. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature Pyramid Networks for Object Detection”, Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2017, pp. 2117–2125.
    [23] S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path Aggregation Network for Instance Segmentation”, Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2018, pp. 8759–8768.
Xem thêm

##plugins.themes.academic_pro.article.sidebar##

Đã Xuất bản

Sep 30, 2025
Download
PDF (English)
Cách trích dẫn
An, N. T., H. H. Hung, và T. T. H. Oanh. “Enhanced YOLOv8 With BiFPN and MHSA for Traffic Vehicle Detection”. Tạp Chí Khoa học Và Công nghệ - Đại học Đà Nẵng, vol 23, số p.h 9D, Tháng Chín 2025, tr 28-32, doi:10.31130/ud-jst.2025.23(9D).558E.

##plugins.themes.academic_pro.article.details##