SCRIBBLE-SUPERVISED CELL SEGMENTATION USING MULTISCALE CONTRASTIVE REGULARIZATION (ISBI 2022)

This is official pytorch implementation of SCRIBBLE-SUPERVISED CELL SEGMENTATION USING MULTISCALE CONTRASTIVE REGULARIZATION (ISBI 2022).

Setup

This work has been implemented on Ubuntu 20.04, Python 3.7, pytorch 1.7.1, CUDA 11.0 and GeForce RTX 3090 GPU.

• Install python packages

pip install -r requirements.txt

• Dataset (BBBC038v1)
  • Download dataset from official website
  • Preprocess code here scribble2label

/bin/bash preprocess_dataset.sh

• Dataset (MoNuSeg)
  • Download dataset from official website
  • Preprocess the dataset

Implementation

• Training

python main.py

• Inference

python inference.py

Citation

Please cite us if you use our work:

@inproceedings{hj_isbi22,
title={Scribble-supervised Cell Segmentation Using Multiscale Contrastive Regularization},
author={Oh, Hyun-Jic and Lee, Kanggeun and Jeong, Won-Ki},
booktitle={2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI)},
pages={},
year={2022},
organization={IEEE}
}