TransWCD: Transformer-based Weakly-Supervised Change Detection

:notebook_with_decorative_cover: Code for Paper: Exploring Effective Priors and Efficient Models for Weakly-Supervised Change Detection [arXiv]

Update

Abastract

Weakly-supervised change detection (WSCD) aims to detect pixel-level changes with only image-level (i.e., scene-level) annotations. We develop TransWCD, a simple yet powerful transformer-based model, showcasing the potential of weakly-supervised learning in change detection.

## :speech_balloon: TransWCD Architectures (Encoder-Only):

## ## A. Preparations ### 1. Download Dataset You can download [WHU-CD](https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html), [DSIFN-CD](https://github.com/GeoZcx/A-deeply-supervised-image-fusion-network-for-change-detection-in-remote-sensing-images/tree/master/dataset), [LEVIR-CD](http://chenhao.in/LEVIR/), and other CD datasets, then use our `data_and_label_processing` to convert these raw change detection datasets into cropped weakly-supervised change detection datasets. Or use the processed weakly-supervised datasets from [`here`](https://drive.google.com/drive/folders/1Ee4T4-pOhZSe9NJ4av4cPBkXh6PX8w71?usp=sharing). Please cite their papers and ours. ``` bash WSCD dataset with image-level labels: ├─A ├─B ├─label ├─imagelevel_labels.npy └─list ``` ### 2. Download Pre-trained Weights Download the pre-trained weights from [SegFormer](https://github.com/NVlabs/SegFormer) and move them to `transwcd/pretrained/`. ### 3.Create and activate conda environment ```bash conda create --name transwcd python=3.6 conda activate transwcd pip install -r requirments.txt ``` ## ## B. Train and Test ```bash # train python train_transwcd.py ``` You can modify the corresponding implementation settings `WHU.yaml`, `LEVIR.yaml`, and `DSIFN.yaml` in `train_transwcd.py` for different datasets. ### ```bash # test python test.py ``` Please remember to modify the corresponding configurations in `test.py`, and the visual results can be found at `transwcd/results/` ## ## C. Performance and Best Models | TransWCD | WHU-CD | LEVIR-CD | DSIFN-CD | |:--------------:|:----------:|:---------------------:|:---------------------:| | Single-Stream | 67.81/[Best model](https://drive.google.com/file/d/1ZK9aNG-RG26ybLGAu9NqaG3-yPChL4Kx/view?usp=drive_link) | 51.06/[Best model](https://drive.google.com/file/d/1z_7e057spJPP4BW_6Ujz8ws-PZ-GWufA/view?usp=drive_link) | 57.28/[Best model](https://drive.google.com/file/d/1i9farsfLxDQBUxxfhjjOJz1zNgjavjoI/view?usp=drive_link) | | Dual-Stream | 68.73/[Best model](https://drive.google.com/file/d/1us1TCqkSfjNjuubasXmRN0vBJ2OAI2vl/view?usp=drive_link) | 62.55/[Best model](https://drive.google.com/file/d/1cpr3ICsR4Ro5XtWA3I8RzDNzxkp1A_Fz/view?usp=drive_link) | 59.13/[Best model](https://drive.google.com/file/d/1oGSlT64WRuzyY1vqmXb1Y55T0VrBJnhn/view?usp=drive_link) | *Average F1 score / Best model On both WHU-CD and LEVIR-CD datasets, the test performance closely matches that of the validation, with differences < 3% F1 score. ## Citation If it's helpful to your research, please kindly cite. Here is an example BibTeX entry: ``` bibtex @article{zhao2023exploring, title={Exploring Effective Priors and Efficient Models for Weakly-Supervised Change Detection}, author={Zhao, Zhenghui and Ru, Lixiang and Wu, Chen}, journal={arXiv preprint arXiv:2307.10853}, year={2023} } ``` ## Acknowledgement Thanks to these brilliant works [BGMix](https://github.com/tsingqguo/bgmix), [ChangeFormer](https://github.com/wgcban/ChangeFormer), and [SegFormer](https://github.com/NVlabs/SegFormer)!