Palmprint identification using radial basis probabilistic neural network

In this paper, we present an effective texture based approach for palmprint recognition. It has three major steps. First, region of interest (ROI) is extracted from the hand image. Then, features are extracted in the processing stage from the palmprint texture using discrete cosine transform (DCT). High frequency coefficients located at the lower right corner of the DCT matrix are omitted considering them as redundant features, thus obtaining dimension reduction. Finally, radial basis probabilistic neural network (RBPNN) is used for classification. It is comparatively a newer method for image classification, comprising features of radial basis function network (RBFN) and probabilistic neural network (PNN). The system is then compared with different ANN classifiers to justify its viability. Average recognition rate of 99.75% demonstrate the efficiency and validity of the method. It's faster in operation, thus making it suitable for online identification as well.