Fingerprint matching algorithm for poor quality images

Abstract

The main aim of this study is to establish an efficient platform for fingerprint matching for low‐quality images. Generally, fingerprint matching approaches use the minutiae points for authentication. However, it is not such a reliable authentication method for low‐quality images. To overcome this problem, the current study proposes a fingerprint matching methodology based on normalised cross‐correlation, which would improve the performance and reduce the miscalculations during authentication. It would decrease the computational complexities. The error rate of the proposed method is 5.4%, which is less than the two‐dimensional (2D) dynamic programming (DP) error rate of 5.6%, while Lee's method produces 5.9% and the combined method has 6.1% error rate. Genuine accept rate at 1% false accept rate is 89.3% but at 0.1% value it is 96.7%, which is higher. The outcome of this study suggests that the proposed methodology has a low error rate with minimum computational effort as compared with existing methods such as Lee's method and 2D DP and the combined method.