Segmentation of slap fingerprint images

Abstract

This paper proposes a novel technique to segment fingerprints from 4-slap image. Black pixels of down scaled and binarized slap images are subjected to experience a mutual superposition attractive force which is inversely proportional to the square of distance. Pixels having high force are grouped based on their force angles, and adjacent connected groups are merged to form components. These components are used to identify hand and to label fingers. The technique have been tested on IITK-Rural and IITK-Student databases and has achieved high segmentation accuracy of 96.80% and 99.2% respectively. © Springer-Verlag Berlin Heidelberg 2013.