Biometric Authentication with a Timeless Learner


One of BATL's precious outcomes is the amount of biometric anti-spoofing data collected. Only in the second phase of the program, two self-test events and two government-controlled-test events were conducted, resulting in tens of thousands of data samples, covering more than a hundred attack species across the fingerprint, face, and iris modalities. Below, we include numbers of samples in each collection event in every modality, and sample data from bona fide and select attack species. The fingerprint collections for Self-Tests 3 and 4 are being released under the name PADISI-USC-Finger. Please, see download and usage instructions here. The face and iris data collected during the self-testing events conducted at USC should be publicly released later this year. Stay tuned!