To address the performance degradation in current motorcycle helmet detection models caused by factors such as large variations in helmet scale, frequent occlusions, and complex backgrounds, this study proposes a motorcycle helmet detection model based on an improved YOLOv11n, which has undergone key improvements in the following areas: First, the C3k2 modules in the original model are partially replaced with C3k2-SCConv modules, where SCConv helps reduce redundant information and enhances the model’s feature extraction capability in complex environments. Second, the iAFF module was introduced to replace the conventional concat operation for feature fusion, effectively leveraging detailed information from shallow layers and semantic information from deep layers, thereby improving the detection performance for small objects. Third, a MultiSEAM module is incorporated into the neck of the model to mitigate information loss caused by occlusion by learning the relationship between occluded and non-occluded regions, which helps reduce missed and false detections owing to occlusion. Finally, ADown modules were used to replace certain convolutional layers, reducing both the number of parameters and computational cost, thereby improving the detection speed. Experimental results demonstrate that the proposed model achieves a 3.4 percentage point improvement in [email protected] compared to the baseline model, while maintaining a competitive detection speed, and overall outperforms existing mainstream object detection models in terms of comprehensive performance.

Tianhang Mu, Lei Ding. Motorcycle Helmet Detection Model Based on Improved YOLOv11n. Academic Journal of Computing & Information Science (2025), Vol. 8, Issue 9: 17-25. https://doi.org/10.25236/AJCIS.2025.080903.

[1] Qin L., Zheng W., Ning P., et al. Awareness and implementation effectiveness of China’s “One Helmet, One Belt” safety campaign among the public[J]. Chinese Journal of Public Health, 2023, 39(09): 1197-1200.

[2] Ross Girshick. Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision. Santiago: NJ: IEEE, 2015:1440-1448.

[3] Wei Liu, Dragomir Anguelov, Dumitru Erhan, et al. SSD: Singleshot multibox detector[C]// Proceedings of European Conference on Computer Vision.Cham. Amsterdam: Springer, 2016: 21-37.

[4] Mingxing Tan, Ruoming Pang, Quoc V Le. Efficientdt:Scalable and efficient object detection[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: NJ: IEEE, 2020: 10781-10790.

[5] Xie P., Cui J., Zhao M. A helmet-wearing detection algorithm for electric bicycle riders based on an improved YOLOv5[J]. Computer Science,2023,50(S1):420-425.

[6] Yuan Y., Tang, W. Helmet-wearing detection for electric bicycle riders based on an improved YOLOv8s model [J]. Journal of Hubei Minzu University (Natural Science Edition), 2024, 42(03): 355-367+367.(in Chinese)

[7] Yang J., Hu P., Dai, J. Helmet-wearing detection algorithm for electric bicycle riders based on YOLOv8-scG neural network [J/OL]. Journal of Chongqing Technology and Business University (Natural Science Edition),2024-07-15.

[8] Zhou S., Peng Z., Zhang H., et al. Helmet-YOLO: A high-precision road safety helmet detection algorithm [J]. Computer Engineering and Applications,2025,61(2):135-144.

[9] Zhou X., Wang K., Zhou X., et al. An improved YOLOv10n-based helmet-wearing detection algorithm for electric bicycles[J]. Electronic Measurement Technology,2025,48(05):40-49.

[10] Khanam R, Hussain M. Yolov11: An Overview of the Key Architectural Enhancements[EB/OL]. https://www.arxiv.org/abs/2410.17725. 2024-10-23.

[11] Jiafeng Li, Ying Wen, Lianghua He.SCConv: Spatial and Channel Reconstruction Convolution for Feature Redundancy[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: NJ: IEEE, 2023:6153-6162.

[12] Yimian Dai, Fabian Gieseke, Stefan Oehmcke,et al.Attentional Feature Fusion[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. Waikoloa: IEEE, 2021: 3560-3569.

[13] Ziping Yu, Hongbo Huang, Weijun Chen, et al.YOLO-FaceV2: A scale and occlusion aware face detector[J].Pattern Recognition, 2024,vol 155:110714.

[14] Wang CY, Yeh IH, Liao HYM. YOLOv9: Learning what you want to learn using programmable gradient information[C]//Proceedings of the 18th European Conference on ComputerVision. Milan: Springer, 2024. 1-21.

[15] Hanhe Lin, Jeremiah D. Deng, Deike Albers, et al.Helmet Use Detection of Tracked Motorcycles Using CNN-Based Multi-Task Learning[J].IEEE Access,2020,vol 8:162073-162084.

[16] Shuo Wang, David C. Anastasiu, Zheng Tang, et al. The 8th AI City Challenge[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: NJ: IEEE, 2024: 7261-7272.