Adaptive Noisy Fingerprint Enhancement Based on Orientation Consistency

Abstract

Fingerprint image enhancement is an essential preprocessing step in fingerprint recognition application. There are many authors proposed the different fingerprint enhancement methods. Most present well-known methods are base on Gabor filters. However, the present proposed Gabor filters almost can not be efficient in some cases such as: the local orientation changes rapidly and the fingerprint image is heavily noisy. This paper proposes a novel filter design technique based on the orientation consistency to improve the Gabor filter with the aim that achieves higher efficiency in clarifying the high curvature ridges of the noisy fingerprint images. It is called Adaptive Orientation Consistency-based Gabor filter (AOC-BGF). Actually, this is the technique which tunes adaptively the Gabor filter window size based on analyzing the orientation consistency. The orientation consistency describes how well the orientations over a neighborhood are consistent with the dominant orientation. The performance of the minutiae detection process on the database FVC2004 DB4 shows the effectiveness and superiority of the proposed method.