Identifying People Using Body Sway in Case of Self-occlusion

Abstract

We propose a method of identifying people in case of self-occlusion by using body sway measured at the head using a top-view camera. To accurately represent the identities of people as reflected in body sway, it is important to acquire accurate appearances in images. However, such images frequently contain defects, especially self-occlusion, that degrade the performance of one of the prevalent methods for identifying people because it uses whole-body regions to identify people. To solve the problem of self-occlusion in this context, our method computes silhouette images of regions at the head by applying a segmentation technique. To reflect people’s identities using body sway, we spatially divide the head region into local blocks and temporally measure movements in them. The results of experiments show that the proposed method can improve the performance of the prevalent method of identification from 17.3% to 57.9%.