Wavelet analysis of cyclic human gait for recognition

There has been increased concern about security during the past few years. Researchers are looking into developing new tools for security enhancement and this has brought biometrics into the limelight. The analysis of human gait as a biometric is relatively newer compared to finger prints, face or iris. This paper presents a new gait feature based on the wavelet analysis of the cyclic gait motion. The proposed feature has low complexity in comparison to the existing techniques and compares well in terms of performance. The new appearance based features are extracted from the silhouettes. The appearance based techniques are generally very sensitive to the quality of the silhouettes. It is observed that the proposed gait recognition technique works better than the current bench mark in case of noisy silhouettes.