https://github.com/zylbuaa/TFormer

The official implementation of "TFormer: A throughout fusion transformer for multi-modal skin lesion diagnosis"
https://github.com/zylbuaa/TFormer

medical-image-analysis multi-modal-fusion skin-lesion-classification transformer

Last synced: 7 days ago
JSON representation

The official implementation of "TFormer: A throughout fusion transformer for multi-modal skin lesion diagnosis"

• Host: GitHub
• URL: https://github.com/zylbuaa/TFormer
• Owner: zylbuaa
• License: mit
• Created: 2023-01-27T10:06:50.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2024-01-29T11:59:17.000Z (10 months ago)
• Last Synced: 2024-08-02T15:32:34.528Z (3 months ago)
• Topics: medical-image-analysis, multi-modal-fusion, skin-lesion-classification, transformer
• Language: Python
• Homepage:
• Size: 521 KB
• Stars: 12
• Watchers: 2
• Forks: 2
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# TFormer

[![MIT Licence](https://badges.frapsoft.com/os/mit/mit.svg?v=103)](https://opensource.org/licenses/mit-license.php)   

## Introduction

The official implementation of "TFormer: A throughout fusion transformer for multi-modal skin lesion diagnosis"

![Our Network Structure](graphical_abstract.png)

## Enviroments

- Windows/Linux both support

- python 3.9

- PyTorch 1.12.1

- torchvision

## Prepare dataset

Please at first download datasets [Derm7pt](https://derm.cs.sfu.ca/Download.html) and then download the pretrained model of swin-tiny on ImageNet-1k from [github](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_tiny_patch4_window7_224.pth). Save the model into the folder "./models/swin_transformer".

## Run details

To train our `TFormer`, run:

```

python train.py --dir_release "your dataset path" --epochs 100 --batch_size 32 --learning_rate 1e-4 --cuda True

```

## License

This project is licensed under the MIT License. See [LICENSE](LICENSE) for details

## Acknowledgement

Our code borrows a lot from:

- [Swin-Transformer](https://github.com/microsoft/Swin-Transformer)

- [Derm7pt](https://github.com/jeremykawahara/derm7pt)