Unsupervised Deep Learning Meets Chan-Vese Model

Dihan Zheng Yau Mathematical Sciences Center, Tsinghua University, China Chenglong Bao Yau Mathematical Sciences Center, Tsinghua University, China; Yanqi Lake Beijing Institute of Mathematical Sciences and Applications, China Zuoqiang Shi Yanqi Lake Beijing Institute of Mathematical Sciences and Applications, China; Department of Mathematical Sciences, Tsinghua University, China Haibin Ling Department of Computer Sciences, Stony Brook University, USA Kaisheng Ma Institute for Interdisciplinary Information Sciences, Tsinghua University, China

TBD mathscidoc:2206.43003

2022.4
The Chan-Vese (CV) model is a classic region-based method in image segmentation. However, its piecewise constant assumption does not always hold for practical applications. Many improvements have been proposed but the issue is still far from well solved. In this work, we propose an unsupervised image segmentation approach that integrates the CV model with deep neural networks, which significantly improves the original CV model's segmentation accuracy. Our basic idea is to apply a deep neural network that maps the image into a latent space to alleviate the violation of the piecewise constant assumption in image space. We formulate this idea under the classic Bayesian framework by approximating the likelihood with an evidence lower bound (ELBO) term while keeping the prior term in the CV model. Thus, our model only needs the input image itself and does not require pre-training from external datasets. Moreover, we extend the idea to multi-phase case and dataset based unsupervised image segmentation. Extensive experiments validate the effectiveness of our model and show that the proposed method is noticeably better than other unsupervised segmentation approaches.
No keywords uploaded!
[ Download ] [ 2022-06-13 21:57:13 uploaded by Baocl ] [ 362 downloads ] [ 0 comments ]
@inproceedings{dihan2022unsupervised,
  title={Unsupervised Deep Learning Meets Chan-Vese Model},
  author={Dihan Zheng, Chenglong Bao, Zuoqiang Shi, Haibin Ling, and Kaisheng Ma},
  url={http://archive.ymsc.tsinghua.edu.cn/pacm_paperurl/20220613215713299346352},
  year={2022},
}
Dihan Zheng, Chenglong Bao, Zuoqiang Shi, Haibin Ling, and Kaisheng Ma. Unsupervised Deep Learning Meets Chan-Vese Model. 2022. http://archive.ymsc.tsinghua.edu.cn/pacm_paperurl/20220613215713299346352.
Please log in for comment!
 
 
Contact us: office-iccm@tsinghua.edu.cn | Copyright Reserved