Vitality detection in personal authentication systems using fingerprints

Fingerprints are considered as the sign of each human being, and this has contributed the development of biometric applications based on such features. Since 2002, an important vulnerability has been shown: it is possible to deceive fingerprint scanners through artificial replicas of fingertips. In order to address this shortcoming it is need to recognize a spoofing attempt with artificial fingers looking for some “life signs” each time an user submit a fingerprint (vitality detection problem). Although this research field is still in its infancy, several methods have been proposed so far, based on additional hardware to the existing capture device and also on fingerprint image processing for extracting those “life sign” from the image captured by the sensor. The first goal of this Ph.D. thesis has been to investigate the current state-of-the-art in fingerprint vitality detection. We compared the performance of each fingerprint vitality detection approach. The second contribution of this Ph.D. thesis is the development of two different new approaches (“power spectrum” and “ridge-width”). The former (in the frequency domain) is based on 2D-Fourier Transform of the fingerprint image. The latter (in the space domain) is based on some morphological considerations (intra-distance ridges and ridge width). Experiments have been carried out on a dataset of images of “live” fingers and fake stamps collected at the DIEE laboratory in Cagliari (82 live fingers and 72 fake stamps, 20 acquisition for each finger/stamp).