Implementation of Gait Recognition for Surveillance Applications

There are various biometric measures that are used in industrial applications for identification of a human. They are signature verification, face recognition method, voice, iris recognition methods, and recognition using digital signatures. These existing human recognition methods have the following limitations of not being unique, low reliability, and could easily traceable by intruders. Gait is the walking style of a human. Gait can be recognized from a view-based approach. In this approach two different image features are required; they are the width of the outer contour of the silhouette and entire binary silhouette. Observation vector can be obtained from the image feature by modeling the frame to exemplar distance (FED) vector sequence with Hidden Markov Model (HMM) as it provides robustness to recognition. In this paper an effort is made for gait recognition useful in real time surveillance applications.