MEDICAL IMAGE RESEARCH · INTERACTIVE DEMO
A closer look.
A clearer workflow.
Follow a real desktop recording, one step at a time. Explore how a segmentation threshold changes what the model highlights.
Open the browser appThe recording starts with an image already loaded. Captions describe the visible actions; empty screen borders have been cropped.
02 · YOUR TURN
From watching
to exploring.
Upload an image or choose a sample. Adjust the threshold, inspect candidate regions, and download the results.
No installation. No account.
Your selected images stay in your browser.
A compact research model
A U-Net model trained on Kvasir-SEG, brought to the browser with the same trained weights.
Explore, then interpret
The overlay shows candidate regions. Changing a threshold changes the selected pixels; it does not establish a diagnosis.
Know the scope
Still-image polyp segmentation only. Inflammation, ulcers, and live video detection are outside this prototype’s scope.