Efficient features extraction for fingerprint classification with multi layer perceptron neural network

In this paper, we present a complete fingerprint recognition system using an Artificial Neural Network (ANN). The ANN is trained by back-propagation algorithm on a set of fingerprint images. Pseudo Zernike Moments (PZM) will be used as a features vector for all images. To detect the region of interest on the fingerprint image, we have used shape information which is characterized by elliptical shape PZM of the elliptical shape constitute the input of the ANN. The data set is divided into two sets, training set and test set. The ANN is trained using the training set and tested on the test set which was hidden during training stage. The recognition rate is measured by the number of correctly classified fingerprints. The structure of the ANN is decided experimentally. The proposed algorithm was tested on a database of more than 400 fingerprints with 10 samples of each person fingerprints. Experimental results have shown that the proposed feature extraction method with an ANN classifer gave a faster training phase and yields a 98% recognition rate.