Project documentation (only Vietnamese): https://vnopenai.github.io/ai-doctor/.
This project contains baseline models for medical image processing and a web user interface for interacting with the system. Our goal is to create an open source code base for students, hobbylists, engineers, or even researchers to get familar with image processing, machine learning and deep learning through medical image processing problems. In this system, we also integrate natural language processing (NLP) models for automatic completion of medical reports.
Want to join us in this project? Send us a message via our contact form.
Click following image to play demo video.
![VN AIDr - Prediction](/VNOpenAI/vn-aidr/raw/master/screenshots/screen.png)
- Python 3.6 + Pip
- Detectron2: https://detectron2.readthedocs.io/en/latest/tutorials/install.html
- NodeJS + Yarn for frontend development
-
Download pretrained models here and extract them into
trained_models
. -
Install requirements:
pip install -r requirements.txt
Or for requirements for CUDA support (require CUDA 10.2):
pip install -r requirements-gpu.txt
-
Update
USE_GPU
inconfigs/common.py
toTrue
to enable GPU support. -
Run the server:
python app.py --port 8080
- Open
http://localhost:8080
in your browser to access web UI.
cd frontend
yarn
yarn start
cd frontend
yarn
yarn build
- Build image
docker build . -t vnaidr
- Run server
docker run -it -p 8080:8080 vnaidr
- Open
http://localhost:8080
in your browser to access web UI.
-
Update
USE_GPU
inconfigs/common.py
toTrue
to enable GPU support. -
Build image
docker build -t vnaidr-gpu -f Dockerfile-gpu .
- Run server
docker run --gpus all -it -p 8080:8080 vnaidr-gpu
- Open
http://localhost:8080
in your browser to access web UI.
The template was developed based on SB Admin 2.