Skip to content

Latest commit

 

History

History
70 lines (45 loc) · 3.46 KB

README.md

File metadata and controls

70 lines (45 loc) · 3.46 KB
title colorFrom colorTo sdk emoji pinned license custom_headers
ohif4hf: Project to showcase AI solutions for medical applications in OHIF
indigo
indigo
docker
🔬
false
mit
cross-origin-embedder-policy cross-origin-opener-policy cross-origin-resource-policy
require-corp
same-origin
cross-origin

ohif4hf

Project to showcase AI solutions for medical applications in OHIF

license CI/CD

ohif4hf was developed by SINTEF Medical Image Analysis to accelerate medical AI research.

The aim of this project, is to enable people without programming expertise to easily test AI solutions on their own data. The AI models should be accessible through a user-friendly user interface (UI), accessible through a browser, not requiring any installation stage.

For the UI, we have used OHIF, which already has all the core features required for reading CTs/MRIs/WSIs and visualizing them, and we will be developing plugins for OHIF, which adds AI model support.

One of the first models we will be adding, are the preoperative tumour segmentation models, which are available in the open software Raidionics.

🤗 Demo

To access the live demo, click on the Hugging Face badge above. Below is a snapshot of the current state of the demo app.

screenshot

NOTE: The project is a work-in-progress. The final plugin(s) are yet to be added. Stay tuned!

The web app is also deployed on my personal website, using HF for hosting the solution.

git clone https://github.com/andreped/ohif4hf.git
docker build -t ohif4hf:latest .
docker run -ti -p 7860:7860 ohif4hf:latest

Then in your favourite browser, go to http://localhost:7860

To go inside docker image and debug, at the bottom of the Dockerfile, add ENTRYPOINT [ "/bin/sh" ] before running.

For development, it might also be useful to build the docker image using --no-cache.

I did not develop OHIF, only implemented some plugins and showcased deployment on Hugging Face space.

Credit should be given to the developers of OHIF for making such an amazing open software solution!

I also want to acknowledge @radames at Hugging Face for assistance with HF space integration.

The code in this repository is released under MIT-License.