A 3D-printed model. Your processor. A piece of ham. AI.
An anonymised attempt on the platform, recorded at UZ Gent.
Three steps from no station to a validated AI score.
Order the 3D-printed HAM model from UZ Gent — ships with all the connections you need. Plug into your scope tower and laptop.
Setup guide →Four progressive shapes, around 15 minutes. Use your unit's standard endoscope, ham, and ink — all sourced locally.
Upload the video. The AI returns speed, stability, time on line, and a percentile against an expert benchmark.
See the methodology →Every attempt produces a short, validated AI report. Speed in mm/s, stability as a percentage of stable frames, time on line, and a percentile that maps to a tercile — trainee, competent, or expert.
Andrea Sorge·Padua
First validation of objective HAM scoring against expert raters in a clinical-grade tip-control task.
Demonstrates stratification across trainees, standard endoscopists, and interventional specialists.
AI scoring shows comparable performance to human raters, enabling scalable assessment.
Build a station this week. Score your first attempt this afternoon.
Open Platform →Questions? Email Lynn Deleu — [email protected]