A novel approach used for measuring fingerprint orientation of arch fingerprint

It is no doubt that fingerprint recognition is the most common biometric modality, which can be used to authenticate the identity of a person. Automatic identification systems based on fingerprint recognition have been extensively deployed in the industrial and forensics area. The performance of these systems relies on the accuracy of the fingerprint comparison algorithm, which is still suffering from the potential displacement and rotation that might occur during fingerprint sample capture process. Especially the strong rotation might decrease the identification accuracy. In order to address the rotation issue, two fingerprint samples can be aligned to each other by analysis of the fingerprint orientation, which is generally represented by the orientation of core point and potentially supported by the connecting line to delta(s). However, for arch fingerprint patterns such a core point doesn´t exist, and some researchers consider the point with maximum curvature as a reference point for an arch fingerprint. But measuring a robust fingerprint orientation for an arch fingerprint is still desirable. In this paper, we propose an approach to measure the fingerprint orientation of arch fingerprints using a set of isosceles triangles. We applied the proposed approach to measure the fingerprint orientation for 80 arch fingerprint samples, which are selected from the challenging CASIA Fingerprint Image database Version 5.0. The experimental results show that the proposed approach is feasible to measure the fingerprint orientation for arch fingerprints, even for a partial imprint with strong rotation.