Speeded-Up Robust Feature extraction and matching for fingerprint recognition

Abstract

Fingerprint recognition refers to the automated method of verifying a match between two human fingerprints which is used to identify individuals and verify their identity. A fingerprint sensor is used to capture a digital image of the fingerprint pattern. The captured image is digitally processed to create a biometric template (a collection of extracted features) which is stored and used for matching. In this paper, we investigate a fingerprint recognition approach by local robust features extraction and matching. In this approach, first the local features are extracted using Speeded-Up Robust Feature (SURF) algorithm. Then the features of the test fingerprint image are compared against two or more exiting template image features for matching. The matching method uses a matching threshold. Two features match when the distance between them is less than the matching threshold. It also eliminates ambiguous matches in addition to using the matching threshold. Finally it calculates the similarity index/matching score from the matching points and take the decision on matching. Since SURF is a scale and rotation invariant algorithm, the fingerprint recognition system shows better recognition accuracy in presence of rotation, scaling and partial distortion of the test image. The experimental results indicate its effectiveness and improved performance over the state of the art.