Biometric system is designed to identify the people accurately, based on human physiological features. The palm print is regarded as one of the most unique, reliable, stable personal charecterstics and performs effectively as a biometric. This paper proposes a new efficient method for the purpose of individual identification using palm print based on Discrete Wavelet Transform (DWT). The wavelet coefficients are used as features for identifying the individual and given to the Support Vector Machine (SVM) classifier for the classification purpose. The wavelet coefficients for palm print are reduced by using specified window size placed at center of the image. Experiments are conducted for various window sizes and the best window size which gives better success rate is chosen. Here using selected window size instead of complete palm print which gives good recognition rate and better result in substantial savings in storage and computation time. Experimental results show promising performance with the proposed methods from NT palm print database.

Key words: Biometrics, Identification, Palm print, Discrete Wavelet Transform, Support Vector Machine.