Benchmark RadarAI BENCHMARK PROFILE
SLAPBench
SLAPBench evaluates multimodal LLMs on four-finger SLAP fingerprint verification using NIST SD302b, with 7,832 image pairs. It assesses identity verification under zero-shot, task-description, and similarity-scoring prompts, reporting metrics like AUC and FAR.
- Released
- 2026-07-17
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
First benchmark for MLLM-based SLAP fingerprint verification, enabling systematic comparison of models and prompting strategies for biometric identity verification tasks.
Motivation
Four-finger SLAP fingerprints are flat live-scan impressions of the index, middle, ring, and little fingers of one hand, used for identity verification in border control and law enforcement.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.