基于SVM的偏瘫患者异常步态识别与临床康复辅助诊断系统

SVM-BASED ABNORMAL GAIT RECOGNITION AND CLINICAL REHABILITATION AUXILIARY DIAGNOSIS SYSTEM FOR HENIPLEGIC PATIENTS

针对脑卒中偏瘫患者的异常步态识别与评估的问题,提出一种基于支持向量机(SVM)的步态分类方法,依据患者下肢行走过程中的连续运动数据对异常步态的细节特征描述,对偏瘫步态进行细分,辅助临床医师对脑卒中患者肢体运动功能异常进行诊断及康复疗效评定。构建穿戴式步态时空参数检测及虚拟现实康复训练系统,提出基于下肢关节角度信息的特征提取方法,建立运动信号与偏瘫步态间的映射关系。基于偏瘫患者在康复治疗中的临床实时步态时空数据,通过对比多种机器学习方法,采用多项式核函数的支持向量机的决策融合模型获得了90%异常步态识别平均准确率,在区分正常与异常步态的基础上,进一步验证了对划圈步态和膝过伸步态的正确诊断。

Aiming at the problem of identification and evaluation of abnormal gait in stroke patients with hemiplegia,this paper proposes a gait classification method based on support vector machine(SVM).According to the detailed characteristics of the abnormal gait based on the continuous movement data of the patient's lower limbs during walking,the hemiplegic gait was subdivided,which assisted clinicians in the diagnosis and rehabilitation of stroke patients with abnormal motor function.A wearable gait spatiotemporal parameter detection and virtual reality rehabilitation training system was constructed,and a feature extraction method based on the angle information of the lower limb joints was proposed to establish the mapping relationship between the motion signal and the hemiplegic gait.Based on the clinical real-time gait spatio-temporal data of patients with hemiplegia in rehabilitation treatment,by comparing various machine learning methods,the decision fusion model of the support vector machine using polynomial kernel function obtained 90% of the average accuracy of abnormal gait recognition.On the basis of distinguishing between normal and abnormal gait,the correct diagnosis of circle gait and knee hyperextension gait was further verified.