Dynamic Attention and Bi-directional Fusion for Safety Helmet Wearing Detection
This provides an incremental improvement for construction safety monitoring by enhancing detection in cluttered environments.
The paper tackled the problem of detecting safety helmet use on construction sites by proposing a novel algorithm with dynamic attention and bidirectional fusion, achieving a 1.7% improvement in mAP@[.5:.95] and reducing GFLOPs by 11.9% compared to baselines.
Ensuring construction site safety requires accurate and real-time detection of workers' safety helmet use, despite challenges posed by cluttered environments, densely populated work areas, and hard-to-detect small or overlapping objects caused by building obstructions. This paper proposes a novel algorithm for safety helmet wearing detection, incorporating a dynamic attention within the detection head to enhance multi-scale perception. The mechanism combines feature-level attention for scale adaptation, spatial attention for spatial localization, and channel attention for task-specific insights, improving small object detection without additional computational overhead. Furthermore, a two-way fusion strategy enables bidirectional information flow, refining feature fusion through adaptive multi-scale weighting, and enhancing recognition of occluded targets. Experimental results demonstrate a 1.7% improvement in mAP@[.5:.95] compared to the best baseline while reducing GFLOPs by 11.9% on larger sizes. The proposed method surpasses existing models, providing an efficient and practical solution for real-world construction safety monitoring.