35 Interactive Medical Imaging Systems Were Never Tested on Real Humans, MICCAI Paper Finds
maier_ak · x · 2026-10-19
A workshop paper at MICCAI 2026 from Johns Hopkins (Sue Min Cho, Catalina Gomez, Mathias Unberath) with 13 co-authors including Katharina Breininger and Nassir Navab calls for human-centred research in medical imaging.
- Narrative keyword review of MICCAI proceedings 2020–2025 (self-described as illustrative, not systematic) surfaced 118 papers involving human evaluation, behavioural data, or user-facing design: 86 in medical image computing, 32 in computer-assisted intervention.
- Sharpest finding: 35 papers propose interactive systems but evaluate them only with simulated users — interactive, never tested with a real person.
- Authors argue interactive segmentation tools, OR navigation displays and prediction explanations succeed or fail in the hands of their intended users, not just on benchmarks; the fix is simple: put the system in front of the people it is for.
More from Research
- Niels Rogge launches paper feed: follow AI research by task, method, author or lab — NielsRogge · 2026-10-19
- A frozen 16MB vocabulary replaces learned embeddings in small and ternary LMs — Sai Kiran Bathula · 2026-10-19
- Meituan LongCat releases RAP: 25K paper-poster pairs as a visual coding benchmark — meituan · 2026-10-19
- BELIEFWEAVER: belief updates and world modeling for long-horizon agent failure recovery — IntelligenceLab · 2026-10-19
- Consistent Distribution Matching: data-free diffusion distillation beats DMD2 with 20GB VRAM — vector-institute · 2026-10-19
- DeepMind's XYEval: one bad user hint drops agent success from 67.4% to 36% — i_dg23 · 2026-10-19