AUTOMATIC DOOR OPENING USING GAIT IDENTIFICATION FOR MOVEMENT AS GESTURE

A wide variety of organizations are using automatedperson identification systems to improve Customersatisfaction, operating efficiency as well as to securecritical resources. Gait identification provides a wayto automatic person identification at distance in visualsurveillance and monitor people without theircooperation. Controlled environments such as banks,military installations and even airports need to beable to quickly detect threats and provide differinglevels of access to different user groups. Gait shows aparticular way or manner of moving on foot and gaitrecognition is the process of identifying an individualby the manner in which they walk. Gait is lessunobtrusive biometric, which offers the possibility toidentify people at a distance, without any interaction orco-operation from the subject; this is the propertywhich makes it so attractive. This paper proposednew method for gait recognition. In this thesis, firststep is extraction of foreground objects i.e. humanand other moving objects from input video sequences orbinary silhouette of a walking person is detected fromeach frame and human detection and tracking will beperformed. After getting binary silhouettes of humanbeings model based approach is used to extractthe gait features of a person. This paper proposesa uncorrelated multilinear discriminant analysis(UMLDA) algorithm for the challenging problemof gait recognition. At last neural network for matlabtool is used for training and testing purpose. Wehave created different model of neural network basedon hidden layer, selection of training algorithm andsetting the different parameter for training. Andthen we will test for the combination of NN+SVM,Knearest neighbour classification. Here all experimentsare done on gait database and input video. Hereall experiments are done on CASIA gait database.Different groups of training and testing dataset givedifferent results. The best recognition result for ourmethod is 96.32%