Palm print recognition

This paper presents a complete study for palm print identification that aims to see how high recognition accuracy can we reach by comparing some results of the previously used line based methods such as Gabor, Canny filters and Modified Finite Radon Transform to represent palm lines and our proposed method that uses basically Radon Transform to describe a person's palm lines. Radon coefficients are used as input features vector, two techniques of classification are used: Matching by correlation and Support Vector Machines. The whole process is applied on two palm print data bases CASIA and PolyU.