Font Size: a A A

Feature Extraction And Recognition Based On The Lower Limbs

Posted on:2013-06-29Degree:MasterType:Thesis
Country:ChinaCandidate:S S BeiFull Text:PDF
GTID:2248330371958450Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
Biometric identification technology is a kind of personal identification method based on the human biological characteristics. The gait recognition is an emerging biometric identification technology, which tends to be realized by the walking posture, i.e. it can realize automatic personal identification by finding and extracting different feature between individuals from the same waling action. Compared with other biometrics feature of fingerprint and face, gait recognition has notable advantages such as long-distance recognizability, difficult imitability, non- aggression and low requirement of clarity. Gait recognition is to analyze the individual gait sequence image acquired by video camera, and realize personal identification with computer technology. Three steps are included, i.e. moving object segmentation, feature extraction, classification and recognition. Firstly, gait sequence image is pre-processed, and background subtraction method is used to detect gait sequence. Secondly, two gait recognition methods are provided, one is based on closed area of low limbs, and the other is based on swing angle of walking feet. The former method is to use the closed area of low limbs as gait feature in each sequence, and process data with wavelet analysis. The latter one is to use swing angle as characteristic parameter, and the angle is got by tiptoe swing with fixed heel as axis. Finally, the more mature BP neural network is used to realize gait recognition. The simulation result of CASIA database shows that gait recognition has identification rate of 90%, which demonstrates this is an efficient identification method.
Keywords/Search Tags:biometrics recognition, gait recognition, wavelet analysis, BP neural network
PDF Full Text Request
Related items