Comparative Test of Smartphone Finger Photo vs. Touch-based Cross-sensor Fingerprint Recognition

摘要

Today’s smart-phone cameras enable touch-less biometric fingerprint capture for practical use. This is inexpensive, comfortable and now fast enough for practical use. This paper deals with the question of how reliably two fingerprints can be compared, when one of them is captured by a smart-phone and the other by a dedicated fingerprint scanner. A multisensor database with 4310 fingerprint images from 108 users was collected. Three touch-based sensors and two touch-less finger photo sensors were used. Latest quality estimations using NFIQ 2.0 and vendor-specific quality metrics complete the study on professional capabilities of finger photo recognition. As side-result of the study, a new touch-less fingerprint processing chain is presented, which reduces the error rates to 1% EER for touch-less- to-touch comparison. This value is comparable to that for touch-to-touch applications.