Structural human shape analysis for modeling and recognition

Abstract

Structural human shape analysis is not a trivial task. This paper presents a novel method for a structural human shape analysis for modeling and recognition using 3D gait signatures computed from 3D data. The 3D data are obtained from a triangulation-based projector-camera system. To begin with, 3D structural human shape data which are composed of representative poses that occur during the gait cycle of a walking human are acquired. By using interpolation of joint positions, static and dynamic gait features are obtained for modeling and recognition. Ultimately, structural human shape analysis is achieved. Representative results demonstrate that the proposed 3D gait signatures based biometrics provides valid results on real-world 3D data. © 2014 Springer-Verlag Berlin Heidelberg.