A novel method for automated human behavior segmentation

摘要: In order to conveniently classify, retrieve, and synthesize human motion, motion capture MoCap data need be properly segmented into distinct behaviors. this paper, we propose a novel automated segmentation method based on posture histograms in sliding window. Firstly, set of new features are proposed defined construct the histogram, which is compact representation behavioral features. Then, by executing window, especially behavior analyzed subsequence level reduce noise sensitivity. We open up way tune window studying steady states behaviors, so that conspicuous stable can obtained. Finally, analyzing clustering property subsequences, problem tactfully simplified detection outlier subsequence. particular, local factor algorithm adopted solve detection, good results achieved. Extensive experiments conducted 14 pieces multi-bcase,CMU Graphics Lab Motion Capture Database: http://mocap.cs.cmu.edu experimental demonstrate our outperforms other state-of-the-art ones. Copyright © 2016 John Wiley & Sons, Ltd.