Genetic Algorithm for Self Occlusion Gait Recognition

Abstract: Human identification has recently gained growing interest from computer vision researchers. Gait recognition domain has a wide range of applications in medicine (biomechanics), mathematics, psychology, rehabilitation and sports activity. It has the unique capability to recognize people at a distance when other biometrics is obscured. This paper aims to handle high levels of occlusion, which is of special importance in gait as the human body is self-occluding. Initially, the silhouettes segmented from video sequences are noisy and need further processing. Mathematical morphological operations are employed for holes remedy and noise elimination. To deal with the incomplete gait image, a new gait pattern representation method named frame difference energy image (FDEI) is used. After the foregoing preprocessing steps, then extract the self occluded feature from the silhouette image sequences. Then, warped template is updated by eliminating self-occluded part. Finally, the genetic algorithm is proposed for recognition of human walking.

Keywords: Gait Recognition, Self Occlusion, Frame Difference Energy Image, Genetic Algorithm.