Human Gait Identification using Depth Gradient Histogram Energy Image

Abstract

Earlier work where we evaluated how traditional Gait Energy Image (GEI) can be used to identify human gait and eventually human ID. However, we also witnessed how inaccurate segmentation (due to unclear foreground-background boundary, human walking direction w.r.t camera axis etc.) may disturb the overall identification performance. To overcome this limitation, present work focuses on using edge and depth gradients extracted from 3D depth data of human gaits. In this paper, first, brief survey of various approaches is outlined. Then promising feature called Depth Gradient Histogram Energy Image (DGHEI) obtained by Histogram of Oriented Gradients (HOG) is presented. Paper also covers performance comparison between 2D GEI and 3D DGHEI based clue to identify human gait.