Utilizing Spatio Temporal Gait Pattern and Quadratic SVM for Gait Recognition

This study aimed to develop a vision-based gait recognition system for person identification. Gait is the soft biometric trait recognizable from low-resolution surveillance videos, where face and other hard biometrics are not even extractable. The cycle pattern of human body locomotion that consists two sequential phases: swing stance. features complete cycle, referred as signature, can be used proposed work utilizes dynamics feature extraction. For this purpose, spatio temporal power spectral utilized captured through sub-pixel motion estimation, they less affected by subject’s appearance. quadratic support vector machine classifier aiming Spatio preserves spatiotemporal adaptable classifier-based across different views appearances. We have evaluated on locally collected dataset captures effect view variance in high scene depth videos. technique achieves significant accuracy all appearances views.