Sweat glands extraction in optical coherence tomography fingerprints

Abstract

Optical coherence tomography (OCT) is a non-invasive technique which can capture high-resolution three-dimension fingerprints. The surface fingerprint, sweat glands and internal fingerprint in OCT fingerprint provides more information than conventional two-dimension fingerprints. In this paper, we present a sweat gland extraction method. The method detects each gland's position using Frangi's filter and segment them by thresholding method. The experiment result on 50 fingers shows our method can successfully segment sweat glands. We also scanned fake fingerprints to do liveness detection. Different from real fingers, fake fingerprints do not have internal structures. We can distinguish fake fingerprints easily with OCT fingerprints. We also show that combining sweat glands information with fingerprint valleys and ridges can improve performance of fingerprint identification.