Skip to content

stegmuel/ScoreNet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ScoreNet

PyTorch implementation of "ScoreNet: Learning Non-Uniform Attention and Augmentation for Transformer-Based Histopathological Image Classification".

Datasets

BRACS

The BRACS dataset is organized as follows:

tree -d BRACS/BRACS_RoI/previous
BRACS/BRACS_RoI/previous
├── test
│   ├── 0_N
│   ├── 1_PB
│   ├── 2_UDH
│   ├── 3_ADH
│   ├── 4_FEA
│   ├── 5_DCIS
│   └── 6_IC
├── train
│   ├── 0_N
│   ├── 1_PB
│   ├── 2_UDH
│   ├── 3_ADH
│   ├── 4_FEA
│   ├── 5_DCIS
│   └── 6_IC
└── val
    ├── 0_N
    ├── 1_PB
    ├── 2_UDH
    ├── 3_ADH
    ├── 4_FEA
    ├── 5_DCIS
    └── 6_IC

Note that to be able to compare with existing baselines we used the "previous" version of the dataset.

BACH

The BACH dataset is organized as follows:

tree -d ICIAR2018_BACH_Challenge/Photos/ 
ICIAR2018_BACH_Challenge/Photos/
├── Benign
├── InSitu
├── Invasive
└── Normal

Acknowledgment

This code relies on some elements of DINO and the accompanying code of Differentiable Patch Selection for Image Recognition.

Cite

@inproceedings{stegmuller2023scorenet,
  title={Scorenet: Learning non-uniform attention and augmentation for transformer-based histopathological image classification},
  author={Stegm{\"u}ller, Thomas and Bozorgtabar, Behzad and Spahr, Antoine and Thiran, Jean-Philippe},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages={6170--6179},
  year={2023}
}