A robust on-line restoration algorithm for fingerprint segmentation

Abstract

Smudged furrows and cut ridges in the image of a fingerprint is the major problem in the fingerprint identification or verification systems. This paper investigates a new on-line unsupervised ridges detection method that reduces the complexity and costs associated with the fingerprint identification procedure. A new detection algorithm for restoration of the ridges of fingerprint is described. This algorithm is based on fuzzy classification technique. The proposed algorithm is compared with several segmentation algorithms. With no cost of time, some new features, such as direction of ridges, have been extracted. The results show a significant improvement in identification performance. Also, the CPU time required for the identification procedure is reduced.