Pytorch implementation of our method for high-resolution HE histology images. This model is based on Generative Adversarial Networks (GAN) that, from an input H&E image, can generate a synthetic Erythroblast Transformation specific related gene (ERG) stained image, highlighting vessel structures.
To use this model, follow the steps bellow.
- Install PyTorch and dependencies from http://pytorch.org
- Install python libraries dominate.
- Clone this repo:
git clone https://github.com/AzmHmd/ERG_Synthesis_model.git
cd ERG_Synthesis_model- Train a model at 1024 x 512 resolution:
python train.py --name [NAME_OF_PROJECT] --dataroot [PATH_TO_DATA] --no_instance
- To view training results, please checkout intermediate results in
./checkpoints/[NAME_OF_PROJECT]/web/index.html.
- Test the model:
python test.py --name [NAME_OF_PROJECT] --dataroot [PATH_TO_DATA] --results_dir [PATH_TO_SAVE] --no_instance
- the trained model is not uploaded due to the size. You can contact the authours to have access to the final generator model after training. Better way to use the trained model is to pull the docker image.
We provide the pre-built Docker image and Dockerfile that can run this code repo. See Dockerfile and get the image by:
docker pull azmhmd/ergsynthesismodel:latest
If you find this useful for your research, please use the following.
@inproceedings{hamidinekoo2021automated,
title={Automated Quantification Of Blood Microvessels In Hematoxylin And Eosin Whole Slide Images},
author={A Hamidinekoo, A Kelsey, N Trahearn, J Selfe, J Shipley, Y Yuan},
booktitle={MICCAI Workshop on Computational Pathology},
year={2021}
}
This code borrows heavily from High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs.