A Preliminary Study on Identifying Fabrication Material From Fake Fingerprint Images

Existing studies suggest the vulnerability of fingerprint verification system against spoof attacks. A spoofing attack occurs when an adversary mimics the biometric trait of another individual for illegitimate access and advantages. Liveness detection algorithms aim to detect live fingerprint samples from the fake artifact. A variety of materials, such as latex, gelatine and silicone, can be used to fabricate fake fingerprint samples. Continuous advancement in the spoofing techniques will lead to the introduction of new materials for fake fingerprint fabrication. However, the performance of these liveness detection algorithms severely degrade when new spoof materials are encountered during the operational stage. Therefore, there is a need for automatic detection of the fabrication material from fake fingerprint images. To this aim, this work investigates texture descriptors such as LBP, BSIF, BGP and GLCM for automatic detection of the fabrication material from fake fingerprint images.