Multiview Gait Recognition Based on Silhouettes Generated after Shadow Detection and Removal Using Photometric Properties Method

摘要

Different biometrics traits are in use nowadays with different accuracy. Gait biometrics is getting popularity among the researchers due to its capability to recognize the people even without cooperation of subject. However, the problems such as shadow detection and shadow removal of moving subjects in visual sur-veillance, less number of images in different conditions, and the data which is rec-orded from surveillance cameras consist of multiple views. This paper presents three algorithms, to remove shadows at the time of silhouette generation from the recorded video, synthetic GEI templates generation to enhance the corresponding gallery probes dataset, and singular value decomposition transformation algorithm to transform the gait feature from one view to another view. Experimental results on a benchmark suite of indoor and outdoor video sequences show that the per-formance of proposed algorithms is better than the other existing algorithms.