作者: Tahirou Djara , Marc Kokou Assogba , Antoine Vianou
DOI: 10.4018/IJCVIP.2016010102
关键词: Mathematics 、 Computer vision 、 Point (geometry) 、 Similarity (geometry) 、 Euclidean distance 、 Artificial intelligence 、 Fingerprint (computing) 、 Matching (graph theory) 、 Pattern recognition 、 Orientation (computer vision) 、 Minutiae 、 Fingerprint Verification Competition
摘要: Most of matching or verification phases fingerprint systems use minutiae types and orientation angle to find matched pairs from the input template fingerprints. Unfortunately, due some non-linear distortions, like excessive pressure fingers twisting during enrollment, this process can cause features be distorted original. The authors are then interested in a method using contactless images for verification. After extraction, they compute Euclidean distances between bifurcation ending points image minutiae. They after ridges angles point orientations. In decision stage, analyze similarity templates. proposed algorithm has been tested on set 420 images. accuracy is found acceptable experimental results promising.