Benchmark RadarReceipt Replay OOD
Receipt Replay OOD is a small out-of-domain benchmark for screen replay detection. It uses receipts, which share planar geometry, curved corners, wear-and-tear artifacts, and text patterns with identity documents, to evaluate document replay detection models under cross-domain conditions without personally identifiable information constraints.
- Released
- 2026-05-26
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Out-of-domain robustness of screen replay detection remains underexplored, especially under realistic domain shifts. This benchmark provides a public dataset for evaluating generalization across domains, which is critical for deployment in varied presentation attack scenarios.
Motivation
Public datasets such as DLC-2021, SynID, and KID34K have significantly contributed to research on presentation attack detection for identity documents, including screen replay attacks.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.