stbya / globalguidance-net

MIA 2021 paper 'Global guidance network for breast lesion segmentation in ultrasound images'.

GlobalGuidance-Net

by Cheng Xue, Lei Zhu, Huazhu Fu, Xiaowei Hu, Xiaomeng Li, Hai Zhang, Pheng Ann Heng

Introduction

This repository is for our MIA 2021 paper 'Global guidance network for breast lesion segmentation in ultrasound images'.

Requirement

• Python
• PyTorch
• torchvision
• numpy
• pydensecrf (here to install)
• training set and testing set (you need to prepare them by yourself and groundtruths should be binary masks)

Training

Train the model with:

python train.py

The pretrained ResNeXt model is ported from the official torch version, using the convertor provided by clcarwin. You can directly download the pretrained model ported by Zijun Deng.

Citation

If this repository is useful for your research, please consider citing:

@article{xue2021global,
  title={Global guidance network for breast lesion segmentation in ultrasound images},
  author={Xue, Cheng and Zhu, Lei and Fu, Huazhu and Hu, Xiaowei and Li, Xiaomeng and Zhang, Hai and Heng, Pheng-Ann},
  journal={Medical image analysis},
  volume={70},
  pages={101989},
  year={2021},
  publisher={Elsevier}
}

Questions

Please contact '[email protected]'