肖军峰1  丁鹏2

（1.中国石油广东销售广州分公司，广东 广州 510000 

2.东华理工大学信息工程学院，江西 南昌 330000）

摘要：膝骨关节炎是导致老年人活动能力受限和身体残疾的主要原因之一，早期发现和干预对于延缓病情发展、改善患者的生活质量具有重要意义。针对现有膝骨关节炎诊断算法检测精度低的问题，提出一种基于改进YOLOv8s的膝骨关节炎自动诊断算法。该算法提出一种改进的卷积块注意力机制模块（CBAM），使网络能够更加关注膝关节图像的关键信息，提高膝骨关节炎的检测精度；设计一种基于多尺度线性注意力的Focal Modula-tion模块，以提高网络的多尺度特征表达能力。实验结果表明，该算法在测试集上的平均精度均值为0.791，有效实现了膝骨关节炎的自动诊断。

关键词：膝骨关节炎；YOLOv8s；Transformer；注意力机制

中图分类号：TP242.2           文献标志码：A          文章编号：1674-2605(2024)06-0004-07

DOI：10.3969/j.issn.1674-2605.2024.06.004                    开放获取

Automatic Diagnosis Algorithm for Knee Osteoarthritis Based on       Improved YOLOv8s

XIAO Junfeng1  DING Peng2  

(1.China National Petroleum Corporation Guangdong Sales Guangzhou Branch, Guangzhou 510000, China

2.School of Information Engineering, East China University of Technology, Nanchang 330000, China)

Abstract: Knee osteoarthritis is one of the main causes of limited mobility and physical disability in the elderly. Early detection and intervention are of great significance for delaying the progression of the disease and improving the quality of life of patients. Aiming at the problem of low detection accuracy of existing knee osteoarthritis diagnosis algorithms, a knee osteoarthritis automatic diagnosis algorithm based on improved YOLOv8s is proposed. This algorithm proposes an improved attention mechanism module (CBAM) of the convolution module, which enable the network to pay more attention to the key information of knee joint images and improve the detection accuracy of knee osteoarthritis; Design a Focal Modulation module based on multi-scale linear attention to improve the multi-scale feature representation ability of the network. The experimental results show that the average accuracy of the algorithm on the test set is 0.791, effectively achieving automatic diagnosis of knee osteoarthritis.

Keywords: knee osteoarthritis; YOLOv8s; Transformer; attention mechanism