install the requirements:
pip install -r requirements.txt
Download all files (weights) from the folder final_weights from the google drive and put them in the folder (../)cell-fine-tune/weights
https://drive.google.com/drive/folders/1Ngp34zO8bx39z0ZSqwtCpZoCjRCEpc24?usp=sharing
Run mask prediction:
python infer.py (--model <hybrid/pure>) (--image_folder <>) (--save_folder <>)
--model hybrid
runs our Hybrid model
--model pure
runs MaskRCNN
Basic statistics can be generated using the notebook: evaluation/evaluate_prediction.ipynb
In case of problems with libomp on Mac (pytorch/pytorch#20030), add:
brew install libomp
https://github.com/Confusezius/unet-lits-2d-pipeline
https://github.com/matterport/Mask_RCNN