Turn Python-based image processing workflows into algorithms that gain extra functionalities.
- Turn your algorithms into web servers and run computations from QuPath, Napari, or Python via HTTP requests.
stardist_blobs.mp4
- Generate dock widgets to run your algorithms interactively in Napari.
oripy_threshold.mp4
- Run your algorithms tile-by-tile on the input data.
tiles.mp4
- Stream results to inspect them in real-time.
yolo-stream.mp4
The documentation is available on this page.
| Task | Examples | Napari | QuPath |
|---|---|---|---|
| Image segmentation | CellPose, StarDist, Instanseg, SAM-2, Rembg, CellPose4 (GPU) | ✅ | ✅ |
| Points detection | Spotiflow | ✅ | ✅ |
| Boxes detection | Ultralytics YOLO | ✅ | ✅ |
| Vectors detection | OrientationPy | ✅ | ✅ |
| Image registration | StackReg, Spam | ✅ | |
| Image denoising | Noise2Void | ✅ | |
| Paths detection | SplineBox | ✅ | |
| Tracking | Trackpy, Trackastra | ✅ | |
| Image generation | Stable Diffusion | ✅ | |
| Live updates | Webcam stream | ✅ |
July 2026
The Imaging Server Kit is being actively developed! Here is what we're up to:
- Developing a bridge with QuPath via
qubalab - Introducing new types (
sk.Any,sk.Plot,sk.Table) - Developing tiled inference via
daskorome-zarr - Integration with
fastapi deploy - Investigating
gRPCfor remote calls
- Mallory Wittwer, EPFL Center for Imaging (mallory.wittwer@epfl.ch)
- Dr. Edward Andò, EPFL Center for Imaging
- Dr. Maud Barthélémy, EPFL Center for Imaging
- Dr. Florian Aymanns, EPFL Center for Imaging
We acknowledge the Personalized Health and Related Technologies (PHRT) initiative for supporting this project.