Skip to content

takimailto/BerDiff

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 

Repository files navigation

BerDiff: Conditional Bernoulli Diffusion Model for Medical Image Segmentation

This is the official implementation of the paper "BerDiff: Conditional Bernoulli Diffusion Model for Medical Image Segmentation" at the 26th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 2023). The pre-print version can be found in arxiv; a camera-ready version will be soon released.

Updates

Dec 2023: initial commit.

Approach

Data Preparation

The LIDC-IDRI dataset can be downloaded from data link. It is preprocessed and given by Probabilistic UNet

Dataset structure:

LIDC/
  |--train/
      |--images/
        |--LIDC-IDRI-0001/
          |--z-105.0_c0.png
          |--z-107.0_c0.png
          ...
        |--LIDC-IDRI-0001/
        ...
      |--gt/
        |--LIDC-IDRI-0001/
          |--z-105.0_c0_l0.png
          |--z-105.0_c0_l1.png
          |--z-105.0_c0_l2.png
          |--z-105.0_c0_l3.png
          |--z-107.0_c0_l0.png
          |--z-107.0_c0_l1.png
          |--z-107.0_c0_l2.png
          |--z-107.0_c0_l3.png
          ...
        |--LIDC-IDRI-0001/
        ...
  |--val/
  |--test/

Checkpoint

A checkpoint can be downloaded from [Google Drive](https://drive.google.com/drive/folders/1HxsR6NJOFILcZzLrcXKfIttctlhdl71I?usp=sharing)

Requirements

- Linux Platform
- torch==1.12.1+cu102 # depends on the CUDA version of your machine
- torchvision==0.13.1+cu102
- Python==3.9.7
- numpy==1.21.2
- blobfile==2.0.0
- argparse==1.4.0

Thanks

Code copied a lot from openai/improved-diffusion

Citation

If you find our work and code helpful, please kindly cite the corresponding paper:

@article{chen2023ascon,
  title={BerDiff: Conditional Bernoulli Diffusion Model for Medical Image Segmentation},
  author={Chen, Tao and Wang, Chenhui and Shan, Hongming},
  journal={MCCAI 2023},
  year={2023}
}

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages