Preprocessing Method of Fingerprint Images in Layered Neural Networks for Individual Identifcation

摘要

We constructed individual identification system of fingerprint in three-layered neural networks, and investigated the effect of fast Fourier transform (FFT) and inverse FFT (IFFT) preprocessing on the performance of individual identification system in layered neural networks. In order to identify individual robustly for degradation of fingerprint images, we suggested the preprocessing method of FFT and inverse FFT (IFFT) in spite of being off-line images. From the results, we found that layered neural networks and the preprocessing method of FFT and IFFT were useful for the identification system with high performance. This may be useful for the feature extraction of fingerprint images