A method for fingerprint enrollment by finger rubbing

Abstract

In these days, most fingerprint enrollment schemes for small mobile sensors ask the user cooperation to acquire wide finger coverage by requesting many input images. Nevertheless, it may still not be enough as some input images may cover the same part of the fingerprint. Therefore, we propose a novel enrollment scheme capturing all input images at one time by rubbing the finger on a sensor without touching-off during the whole enrollment process. Then, optimal image selection is followed for selecting best images among all input images to maximize the fingerprint coverage with less number of enrollment images. Experimental results showed that Equal Error Rate (EER) is about 30% improved compared to the conventional enrollment scheme.