Benchmark RadarHounsBench
HounsBench is a CT-centric patient-state benchmark evaluating three task families: readout, reconstruction, and simulation. It provides patient-disjoint splits and per-family metrics for evaluating models on volumetric medical images and clinical language.
- Released
- 2026-08-13
- Readiness
- Runnable
- Primary field
- Health & Life Sciences
Why it matters
HounsBench addresses the evaluation gap for CT-centered intelligence by unifying diverse tasks under a shared patient-state inference framework, enabling assessment of models that must integrate imaging and language for clinical decision support.
Motivation
Clinical intelligence requires estimating a patient's underlying condition from incomplete observations rather than learning isolated mappings from scans to answers.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.