Authors : M. Manju and V. Kavitha
Abstract: In this study, the accuracy of finger print recognition problem is been addressed. As per the literature the Back Propagation Network (BPN) for fingerprint recognition has resulted in inconsistent with unpredictable performance. This research has proposed the soft computing tool to images to overcome the low recognition rate and the low accuracy in fingerprint identification. Fuzzy logic is worn to eliminate the false minutiae from the fingerprint. Genetic algorithm has been incorporated to optimize the weights of neural network and the accuracy in the recognition process has been improved. The proposed method is implemented on the FVC 2004 DB1 database. The Laplacian based Pyramidal Model has strongly supported in fingerprint enhancement process which has increased the Peak Signal to Noise Ratio (PSNR) and decreased the Mean Square Error (MSE). The results have proven that the false minutiae have been eliminated by applying fuzzy rules and also the Equal Error Rate (EER) has been reduced. The increase in the recognition accuracy moreover in turn has reduced the training and the testing time.
M. Manju and V. Kavitha , 2013. Fuzzy Rule Based Neuro-Genetic Approach for Fingerprint Recognition. International Journal of Soft Computing, 8: 154-162.