Gait Based Recognition

Gait recognition deals with recognizing an individual uniquely based on the ways/he walks. Recently a lot of attention in computer vision has been devoted towards gait recognition because it has important applications in surveillance and biometric authentication. This thesis discusses different methods to solve the problem of gait recognition using joint angles, model and parameter estimation, shape and optical flow and methods that use entire images to do gait based recognition i.e. based on global features. We provide basis for the development of an optimal gait recognition system and finally we evaluate some challenges that still need to be solved and conclude with some thoughts about future directions for human gait recognition