Academic Journal of Computing & Information Science, 2024, 7(10); doi: 10.25236/AJCIS.2024.071010.
Shouhao Zheng1, Feng Tian2
1College of Intelligent and Information Engineering, Shandong University of Traditional Chinese Medicine, Jinan, 250399, China
2School of Data Intelligence, Yantai Institute of Science and Technology, Yantai, 265600, China
Aiming at the problems of low detection accuracy and low detection efficiency of Yolov8s algorithm, an improved Yolov8n algorithm with high accuracy and few parameters is proposed. Specifically, to address the problem of a large number of parameters in the basic Yolov8s, its variant Yolov8n algorithm is introduced to improve the model training efficiency. Subsequently, for the problem that the original CIOU function cannot be accurately localised, the WIOU loss function is introduced to consider the degree of overlap between different regions and accurately mask the boundaries, which further improves the segmentation performance. Finally, an SGD optimiser is introduced on this basis to improve the model training efficiency. The improved Yolov8n algorithm is applied to the VOC2007 dataset for training and validation. The experimental results show that the proposed improved Yolov8n algorithm has higher training accuracy and fewer parameters with high training efficiency, reaching 85.6% for P, 76.9% for R and 83.3% for map on the dataset, which has high detection accuracy and validates the model. Compared with other models, the number of parameters is reduced, and the detected maps are improved by 0.6% ~ 17.1% and significantly improved on several detection categories, which verifies the superiority of the model. In summary, the improved yolov8n algorithm proposed in this paper can meet the functional requirements of high accuracy and high efficiency of traffic object detection in road traffic equipment as well as self-driving traffic roads.
Yolov8n, WIOU, SGD, VOC2007
Shouhao Zheng, Feng Tian. An optimized Yolov8n approach and its application to object detection problems. Academic Journal of Computing & Information Science (2024), Vol. 7, Issue 10: 67-75. https://doi.org/10.25236/AJCIS.2024.071010.
[1] Jiang H. Overview of autopilot technology in industry[J]. Logistics technology and application, 2024, 29(03): 106-107.
[2] Kamal U, Tonmoy T I, Das S, et al. Automatic traffic sign detection and recognition using SegU-Net and a modified Tversky loss function with L1-constraint[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 21(4): 1467-1479.
[3] Girshick R, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[J]. Computer Vision and Pattern Recognition , 2014(1): 580-587.
[4] Girshick R. Fast R-CNN[C]. Proceedings of the IEEE international conference on computer vosion, 2015: 1440-1448.
[5] Liu X, Cheng J, Cheng Y, et al. Distance posts detection and character sequences recognition method in video images acquired from camer in moving vehicle[J]. Computer Engineering and Applications, 2023, 59(8): 175-81.
[6] Liu H, Li J. Vehicle type detection algorithm based on YOLOv5s model[J]. Journal of Lang fang Normal College (Natural Science Edition), 2022, 22(3): 24-28.
[7] Li S, Shi T, Jing F. Improved road damage detection algorithm for YOLOv8[J]. Computer Engineering and Application, 2023, 59(23): 165-174.
[8] Kingma D, Ba J. Adam: A method for stochastic optimization in: Proceedings of the 3rd international conference for learning representations (iclr'15)[J]. SanDieg0, 2015, 500.
[9] Wang S, Ton W, Huang J, et al. Tile surface defect detection based on YOLOv5[J] . Packaging Engineering, 2022, 43(09): 217-224.
[10] Sun Y, Li W, Zhou N. Study on improved YOLOv5m method for cargo category detection[J]. Computer Engineering and Applications, 2023, 59(16): 205-211.
[11] Shang C, Wang M, Ning J, et al. Identification of high similarity dairy goats under joint loss optimization[J]. Chinese Journal of Image and Graphics, 2022, 27(04): 1137-1147.
[12] Safaldin M, Zaghden N, Mejdoub M. Moving object detection based on enhanced Yolo-V2 model[C]. IEEE, 2023: 1-8.
[13] Wang C, Bochkovskiy A, Liao H. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]. Vancouver: IEEE, 2023: 7464-7475.
[14] Lin T, Dollar P, Girshck R, et al. Feature pyramid networks for object detection[C]. Proceedings of the IEEE: 2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), IEEE, 2017: 936-944.
[15] Liu S, Qi L, Qin H, et al. Path aggregation network for instance segmentation[C]. Proceedings of the IEEE: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). IEEE, 2018: 8759-8768.
[16] Zhang Y, Ren W, Zhang Z, et al. Focal and efficient IoU loss for accurate bounding box regression [J]. Neurocomputing, 2022, 506: 146-157.
[17] Zheng Z, Wang P, Ren D, et al. Enhancing geometric factors in model learning and inference for object detection and instance segmentation[J]. IEEE Trans-actions on Cybernetics, 2022, 52(8): 8574-8586.
[18] Li M, Xiao Y, Zong W, et al. A lightweight chestnut fruit recognition method based on improved Y0L0v8 model[J]. Journal of Agricultural Engineering, 2024, 40(1): 201-209.
[19] Jian W, Gauri J. Cooperative SGD: A unified framework for the design and analysis of local-update SGD algorithms[J]. Journal of Machine Learning Research, 2021, 22(1): 213-223.
[20] Evringham M, Van C, Williams C, et al. The pascal visual object classes (voc) challenge[J]. International journal of computer vision, 2010, 88(2): 303-338.