Adaptive and Faster Approach to Fingerprint Minutiae Extraction and Validation

摘要

Fingerprint has remained a very vital index in the field of security where series of Automatic Fingerprint Identification System (AFIS) have been developed for human identification. Many of these systems involve matching each of the features of a template image with each of the features in the feature sets in the reference database to determine the level of match between the template and the reference images. Matching is done on the basis of preset parameters such as feature type, location, orientation and so on. Obtaining the features from the template image and for building a reference database involves the implementation of a sound fingerprint feature detection and extraction algorithm. In this paper, the process of detecting and extracting false and valid features contained in a fingerprint image is discussed. Some of the existing fingerprint features extraction algorithms were firstly modified and the resulting algorithms were implemented. The implementation was carried out in an environment characterized by Window Vista Home Basic as platform and Matrix Laboratory (MatLab) as frontend engine. Fingerprints images of different qualities obtained from the manual (ink and paper) and electronic (fingerprint scanner) methods were used to test the adequacy of the resulting algorithms. The results obtained show that only valid and true minutiae points were extracted from the images.