YOLOv8-Based Hand Bone Fracture Image Detection Algorithm

在医疗领域，手骨骨折图像检测具有极其重要的实际应用价值，尤其是在面对复杂的骨折情况时，精准识别手骨骨折目标依然是一个亟待解决的难题。针对这一问题，本文提出了一种基于YOLOv8算法的改进方法，通过引入CBAM注意力机制，有效增强了网络对不同尺度特征的关注能力，进而显著提升了在小目标以及复杂场景下的检测精度。实验结果清晰地表明，经过改进的YOLOv8模型相较于原YOLOv8模型，其平均精度均值实现了0.57个百分点的提升，展现出对手骨骨折图像更为出色的检测效果。Hand bone fracture image detection holds extremely significant practical application value in the medical field. Especially when confronted with complex fracture situations, accurately identifying hand bone fracture targets remains a pressing challenge that needs to be addressed. To tackle this issue, this paper proposes an improved method based on the YOLOv8 algorithm. By introducing the CBAM attention mechanism, the network’s ability to focus on features of different scales is effectively enhanced, thereby significantly improving the detection accuracy for small targets and complex scenes. The experimental results clearly demonstrate that the improved YOLOv8 model, compared with the original YOLOv8 model, has achieved a 0.57 percentage point increase in mean average precision (mAP), showing a more outstanding detection effect on hand bone fracture images.