Comparative Study of Fingerprints Liveness Detection Using Noise in Ridge Valley Structure, Texture Analysis, and CNN Method

Fingerprint is a more reliable and distinctive biometric. The usage of it in security applications is widespread. However, it is simple to spoof fingerprints obtained from scanners by creating false fingers out of gelatin, fevicol, clay, silicon rubber, play-dough, and moldable plastic. In these situations, it is crucial for the researcher to be able to differentiate among real and prosthetic fingers. The liveness detection of fingerprints is a technique to identify actual fingers acquired by fingerprint scanners for person identification by detecting liveness/physiological indications from fingerprints. Convolutional neural networks (CNNs) were employed in this study to decide the liveness of a fingerprint. When compared to the best previously reported findings, our top model obtains total rate of 95% of properly categorized samples.