Active Contour Loss

Implementation of active contour loss function for medical image segmentation based on "Learning Active Contour Models for Medical Image Segmentation" by Chen, Xu, et al.

Introduction

==The arXiv version of this paper will be available soon. ==

Requirements

Tensorflow >= 1.5

Keras >= 2.0

Numpy

Training

A pretrained model might be suggested to use, because somtimes active contour loss function may not be stable in the early steps for training.

Citation

If you find Active-Contour-Loss is useful in your research, please consider to cite:

@inproceedings{chen2019learning,
  title={Learning Active Contour Models for Medical Image Segmentation},
  author={Chen, Xu and Williams, Bryan M and Vallabhaneni, Srinivasa R and Czanner, Gabriela and Williams, Rachel and Zheng, Yalin},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  pages={11632--11640},
  year={2019}
}

Other Re-implementation

...