Segmentation of fingerprint image by using polarimetric feature

Abstract

Segmentation of fingerprint image is the first step of fingerprint recognition, and it plays an essential role which helps to preserve genuine and reduce false minutiae and further aids the performance of Automatic Fingerprint Identification System (AFIS). The problem of segmentation has been thoroughly studied but never been completely solved. During this paper, we propose a novel representation of fingerprint image with the added polarimetric information which is captured by Stokes Imaging System, and followed by a simple yet efficient segmentation of fingerprint image based on the polarimetric variance (Polvar). Polarimetric characteristic is another distinguishable feature beside intense that reflecting light carries, and it provides potential way to enhance the contrast between background and foreground, and between ridges and valleys as well. And therefore, there is a possibility to achieve a satisfactory segmentation results. Nonoverlapping block Polvar feature is utilized to accelerate computation, and moreover the segmentation results that are based on other common used features are compared, that is block energy, block coherence, block cluster degree. Experimental results show that our proposed novel method is much efficient than the other features and simultaneously achieve higher accuracy especially it well segments the case of remaining ridges from previously scanned finger. Segmentation results are evaluated both visually by human inspire and quantitatively.

Two-Level Evaluation on Sensor Interoperability of Features in Fingerprint Image Segmentation

“…), texture features [ 3 , 5 , 7 , 9 , 19 , 20 ] (such as gradient, coherence, Gabor response, etc . ), and other features [ 12 , 21 – 23 ] (such as Harris corner point features, polarimetric feature, number of invalidated minutiae, etc .).…”