摘要
针对背包、外套等干扰因素致使步态识别率降低的问题,提出一种分割加权的步态识别方法。首先通过帧差法与阈值分割相结合的方法得到动态信息更加丰富的帧差阈值能量图(FTDEI),将该能量图分割为3部分,并对每部分添加相应权重,然后利用Gabor小波对分割加权后的FTDEI进行不同角度的特征提取,得到加权的滤波特征(SWA-Gabor),最后通过KNN分类器对SWA-Gabor特征进行分类和识别。基于分割加权的步态识别方法能够很好地避免背包等干扰因素的影响。为了验证该算法的识别效果,在中国科学院自动化研究所CASIA-B步态数据库上进行实验,结果表明,在携带背包和外套的情况下,该算法的识别率较其它算法提高了约5%,取得了很好的识别效果。
In order to solve the problem that the gait recognition rate is decreased due to the interference factors such as backpacks and jackets,a segmentation weighted gait recognition method is proposed.Firstly,a frame difference threshold energy map（FTDEI）is obtained by combining frame difference method with threshold segmentation,which obtains richer dynamic information We divide the energy map into three parts,add corresponding weights to each part,and use Gabor wavelet to extract the characteristics of FTDEI from different angles.Weighted filtering features（SWA-Gabor）are obtained.Finally,SWA-Gabor features are classified and identified by KNN classifier.The gait recognition method based on segmentation weighting can reduce the influence of interference factors such as backpacks by weighting.In order to verify the recognition effect of the proposed algorithm,experiments are carried out on CASIA-B gait database.Experimental results show that the recognition rate of the proposed algorithm is improved by about 5 % compared with other algorithms in the case of carrying backpacks and outerwear.
作者
郑慧平
李旭健
张阳
ZHENG Hui-ping;LI Xu-jian;ZHANG Yang(College of Computer Science and Engineering,Shandong University of Science and Technology,Qingdao 266590,China)
出处
《软件导刊》
2018年第7期40-43,共4页
Software Guide
