This project is a proof of concept, where we use a 3D printed Openflexure microscope to take pictures and then detect malaria. We use NIH images to train our model.
The app also functions without an openflexure microscope just by uploading images.
Run the docker file from a laptop/ computer while the Openflexure microscope is up and running. First build the dockerfile from within the malatec_app/docker folder:
docker build -t malatec_app .
# if you are using an AArch64/ARM64 CPU (e.g. a Raspberry Pi with a 64-bit operating system), use:
docker build -f Dockerfile.aarch64 -t malatec_app .Now you can run the docker file, which starts up the streamlit app:
# -p 8501:8501 is needed if you want to serve other computers in your network,
# be sure to use the external IP address of the docker host (not the docker container) to access streamlit in a browser
docker run --rm --name malatec_app -p 8501:8501 malatec_app:latestFollow the link provided in the console and you should be able to see the streamlit app.
Pipeline
- Cell segmentation from microscope images using Article | GitHub
- Classification: SqueezeNet + VGG19 Article | GitHub
- Powered by PyTorch, Tensorflow and Streamlit
Publicly Available Malria Datasets
Made with ❤️ in Switzerland ⛰️