Reads numerical readouts (device parameters, timings, settings) directly off ophthalmological surgical video — useful when raw device telemetry isn't recorded but the screen is. Combines region detection with a small recognizer trained on the fonts surgical equipment actually uses. TODO: refine with real accuracy numbers, the dataset story, and a frame-level demo gif.
SurgeryOCR
activeOCR for ophthalmological surgical video overlays.
visionocrsurgerypython
⊟
visualization tbd
drop a hero image at
public/projects/surgery-ocr.png and set image: "/projects/surgery-ocr.png" on this project in content/projects.ts