3D Human Motion Analysis for Reconstruction and Recognition

In recent years, biometrics modalities with depth information are an interesting resource. As they can apply to many applications, range scanners have obviously become popular increasing the measurement accuracy and speed. In this paper, we propose a method for 3D human motion analysis for reconstruction and recognition. We use 3D gait signatures computed from 3D data that are obtained from a triangulation-based projector-camera system. The method consists of several steps: First, 3D human body data are acquired by using a projector-camera system. The body data are composed of representative poses that occur during the gait cycle of a walking human. Second, 3D human body model is fitted to the body data using a bottom-up approach to estimate its pose. Third, the entire gait sequence is recovered by interpolation of joint positions in the fitted body models. Representative results have been shown to ensure the robustness of the proposed method.