The TFPA database is a database consisting of 1,608 T-shirt attacks using 100 unique presentation attack instruments (PAIs). The T-shirt PAIs are based on synthetically generated face images, each worn by at least two subjects across eight different capturing scenarios and two capturing streams (visible + depth).
Download and Copyright:
The database is being made available for researchers from 2023 onwards. Interested researchers can download this database upon request. Please send an email to mathias.ibsen@h-da.de. Any commercial use/distribution of this database is prohibited. All the technical reports and papers that report experimental results from this database should provide acknowledgment and reference to [1].
References
[1] M. Ibsen, C. Rathgeb, F. Brechtel, R. Klepp, K. Pöppelmann, A. George, S. Marcel, C. Busch, „Attacking Face Recognition with T-shirts: Database, Vulnerability Assessment and Detection“, in IEEE Access, 2023.