OHIF Viewer: Clinical-Grade Medical Imaging Running on WebGL in Your Browser

Most WebGL projects you see here are about beauty, play, or spectacle. This one is about looking inside people. The OHIF Viewer is a zero-footprint medical imaging platform that runs entirely in your browser. No installs, no plugins. Point it at a DICOM archive and suddenly you're scrolling through CT slices, fusing PET and MRI volumes, drawing segmentation masks on tumors. The kind of software radiologists use daily to make decisions that actually matter. It loads fast because it streams pixel data progressively, fetching metadata first and pulling image frames on demand.

The project is maintained by the Open Health Imaging Foundation, a program of Massachusetts General Hospital, with Radical Imaging leading architecture and development. Kitware's VTK.js handles the heavy 3D visualization. The full source lives at github.com/OHIF/Viewers, and the rendering engine, Cornerstone3D, is worth studying on its own.

Under the hood, the trick that makes everything work is Cornerstone3D's offscreen rendering approach. Instead of spinning up a separate WebGL context per viewport (which would hit browser limits fast in a layout with ten panels), it drives every viewport from a single shared WebGL context, rendering offscreen and compositing the results. WebAssembly handles JPEG 2000 and JPEG-LS decompression in parallel via web workers. Annotations render as SVG overlays, so they stay sharp at any zoom. The whole thing is React-based with a mode/extension architecture that lets you swap out entire clinical workflows without touching core rendering code. If you've ever wondered what a serious, production-grade WebGL application looks like when the stakes are higher than frame rate, dig into the Cornerstone3D source and the OHIF docs. It's one of the most sophisticated uses of GPU-accelerated rendering on the web, and it's saving lives while it's at it.