Processing of Distorted Fingerprints with use of Three-Rate Hybrid Kohonen Neural Networks
Igor Astrov, Svetlana Tatarly, Sergei Tatarly, Ennu Rüstern · 2006
In this paper, we present the three-rate hybrid Kohonen neural network (TRHKNN) for image processing. The received TRHKNN consists of "fast" Kohonen neural network (FKNN), "middle" Kohonen neural network (MKNN) and "slow" Kohonen neural network (SKNN). The illustrative example - recognition and restoration of distorted fingerprint - was carried out using the TRHKNN. The received TRHKNN has not only high speed of image recognition, but also high speed of image restoration. This example demonstrates that the proposed TRHKNN is capable not only to identify the distorted image of fingerprint but also to restore the undistorted image of fingerprint. The simulation results with use the software package Simulink show the computing procedure and applicability of TRHKNNs for fast-acting image processing and analysis in real-time conditions