WEAKLY SUPERVISED NUCLEI SEGMENTATION VIA INSTANCE LEARNING

Description

This page contains the code of "WEAKLY SUPERVISED NUCLEI SEGMENTATION VIA INSTANCE LEARNING". This work has been accepted by IEEE International Symposium on Biomedical Imaging (ISBI), 2022 as Oral Presentation. For more details, please refer to https://arxiv.org/abs/2202.01564.

Run the code

0. Create environment under CUDA 10.2

conda create -n nucseg python=3.7 h5py
conda activate nucseg
pip install torch==1.4.0 torchvision==0.5.0 -f https://download.pytorch.org/whl/cu102/torch_stable.html
pip install -r requirements.txt

1. Train SPN

python main.py --id SPN --cfg network/exp/MO/SPN.yaml --gpu 1

2. Train IEN

python main.py --id IEN --cfg network/exp/MO/IEN.yaml --gpu 1

3. Model inference

python main.py --id IEN_infer --cfg network/exp/MO/IEN_infer.yaml --gpu 1

Citation

If you find this code helpful, please cite our work:

@article{liu2022weakly,
  title={Weakly Supervised Nuclei Segmentation via Instance Learning},
  author={Liu, Weizhen and He, Qian and He, Xuming},
  journal={arXiv preprint arXiv:2202.01564},
  year={2022}
}