基于密集光流的步态识别

删除或更新信息，请邮件至freekaoyan#163.com(#换成@)

本站小编 Free考研考试/2024-01-16

-->

商磊,张宇,李平.基于密集光流的步态识别[J].,2016,56(2):214-220

基于密集光流的步态识别

Dense optical flow-based gait recognition

DOI：10.7511/dllgxb201602016

中文关键词:步态识别背景减除密集光流降维

英文关键词:gait recognition background subtraction dense optical flow dimensionality reduction

基金项目:国家自然科学基金青年基金资助项目(61005085)；中央高校基本科研业务费专项资金资助项目(2012QNA4024).

作者	单位
商磊,张宇,李平

摘要点击次数:1280

全文下载次数:2088

中文摘要:

作为一种生物特征，步态在视频监控、行为分析等领域具有很大的应用前景．提取步态特征的关键在于对步态在时间、空间两个维度上的变化模式进行描述．基于密集光流提出了一种步态特征提取算法，通过密集光流表征每帧图像人体区域各部位的运动强度和方向，综合一个步态周期内所有单帧特征作为步态周期的特征．利用主成分分析、线性判别分析对步态特征进行降维处理，用支持向量机进行分类，验证提取特征的分类性能．实验结果表明，所提算法采用光流特征，提供了丰富的动态信息，可以很好地描述步态在时间维度上的变化，在与现有步态特征描述算法的对比中，体现出了良好的识别性能．

英文摘要:

As a biological feature, gait has great application prospect in many fields, such as video surveillance, behavior analysis and so on. The key point of extracting gait feature is to describe the change pattern of gait both in spatial and temporal dimensions. The proposed gait feature extracting algorithm is based on dense optical flow. Dense optical flow can offer the intensity and orientation of human motion of each point in the subject region, which is the feature of a single frame. Synthesizing the features of each frame in one gait cycle can obtain the feature of a gait cycle. The dimension of gait feature is reduced by principal component analysis (PCA) and linear discrimination analysis (LDA). Then, the subjects are classified by support vector machine (SVM) to verify the classification ability of the extracted feature. The experimental results show that the proposed algorithm uses the optical flow feature to offer rich dynamic information, which can describe gait′s change in temporal dimension well, and is proved to have a better recognition performance compared with the other gait representations.

查看全文查看/发表评论下载PDF阅读器

关闭

基于密集光流的步态识别

本站小编 Free考研考试/2024-01-16

相关话题/

领限时大额优惠券,享本站正版考研考试资料!