Palmprint recognition using 2d-gabor wavelet based sparse coding and RBPNN classifier

This paper proposed a novel and successful method for recognizing palmprint using 2D-Gabor wavelet filter based sparse coding (SC) algorithm and the radial basis probabilistic neural network (RBPNN) classifier proposed by us Features of Palmprint images are extracted by this SC algorithm, which exploits feature coefficients' Kurtosis as the maximum sparse measure criterion and a variance term of sparse coefficients as the fixed information capacity At the same time, in order to reduce the iteration time, features of 2D-Gabor wavelet filter are also used as the initialization feature matrix The RBPNN classifier is trained by the orthogonal least square (OLS) algorithm and its structure is optimized by the recursive OLS algorithm (ROLSA) Experimental results show that this SC algorithm is successful in extracting features of palmprint images, and the RBPNN model achieves higher recognition rate and better classification efficiency with other usual classifiers.