Fast pore matching method based on core point alignment and orientation

Nowadays, high-resolution fingerprint images are more and more used in the fingerprint recognition systems thanks to the recognition accuracy that they provide. Indeed, they offer more sufficient details such as sweat pores, ridges, contours, and other details. Pores have been adopted to be one of the brilliant nominees in improving the efficiency of automated fingerprint identification systems to maintain a high level of security. However, the geometric transformations, that occur during the acquisition phase, can cause several defects on the result of the matching process, hence they decline the accuracy of the recognition. To overcome this problem, alignment is often needed. This image pretreatment is classically based on complex geometric operations that are time-consuming. Otherwise, for pore matching, the majority of approaches are based only on pore coordinates. In this paper, we propose a novel pore matching method based, firstly, on only one of the singular points, namely the core points for the alignment phase, and also the valuable features used for the score calculation namely position and the orientation of pores. We assess our proposed approach using the PolyU-HRF database and we compare it to some well-known approaches of level 3 fingerprint recognition. The experimental results demonstrate that the proposed method can achieve significant performance recognition accuracy across various qualities of fingerprint images.