An Advanced Multi Feature based Gait Recognition

Saeid Fazli and Hadis Askarifar


Gait recognition, Human motion analysis, Principal Component Analysis (PCA), Biometrics


Person recognition based on the way he/she walks has great interest amongst computer vision researchers. Gait recognition has been an active research topic in recent years [1]. This paper closely focuses on three steps of gait recognition including preprocessing, feature extraction and classification. In the first step, we produce a standardized dataset. In the second step, an OR- contour based multi feature system is used and in the last step, we apply a neural network classifier. NLPR (National Laboratory of Pattern Recognition Institute of Automation Chinese Academy of Sciences) databases are used in our experiments. Experimental results show the effectiveness of our proposed method in gait recognition. The results are also compared to other existing methods.

Important Links:

Go Back