Person Re-identification Using 3D Data Analysis Method and Kinect Sensor

Abstract

Automated personal identification systems, such as personal facial recognition systems and automated motor vehicle registration number checking, are examples of public security protection systems. The area of personal identification for security purposes is of growing interest for security assessment in public places, and airports, as examples, now becoming an imperative matter for research in the Internet netscape. We propose a method of immediate recognition of a subject person, based on Incremental Dynamic Time Warping (IDTW) which identifies personal gait patterns recorded via a 3D depth sensing camera such as in Microsoft’s Kinect® version 2, by analyzing a dataset of gait gestures derived from a sample of 16 people. The experimental results show that the IDTW algorithm increases the efficiency of recognizing at 81%.