Human Identification by Gait Recognition

As a biometric, gait has several attractive properties.Acquisition of images of an individual’s gait can be done easily inpublic areas, with simple instrumentation, and does not require thecooperation or even awareness of the individual under observation.Recognition of gait through gait analysis is an important researchtopic, with potential application in video surveillances, tracking,access control, smart interfaces and monitoring. Gait recognition is theprocess of identifying an individual by the manner in which they walk.The aim of this paper is the to compare the two algorithms of gaitrecognition system. There are so many methods like dynamic timewarping, gait energy image, hidden markov method, etc. is explainedby authors. In this paper we proposed an automatic gait recognitionapproach based on the features like variance and edge of videosequences. The gait signature vectors are constructed to identifydifferent subjects. Finally, similarity measurement based on the k NNclassifier is carried out to recognize the different subjects in whichthey walk. Finally, feature outputs for variance and edge detection arecompared. The results obtained by variance method and edge detectionmethod are 94% and 68.8% resp. From graphical output we concludedthat the results obtained by variance methods give the better outputthan canny edge detection method.