# DANet

**Repository Path**: zhangzexu/DANet

## Basic Information

- **Project Name**: DANet
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-11-03
- **Last Updated**: 2021-11-03

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# [Dual Attention Network for Scene Segmentation(CVPR2019)](https://arxiv.org/pdf/1809.02983.pdf)

[Jun Fu](https://scholar.google.com/citations?user=h3vzrgkAAAAJ&hl=zh-CN), [Jing Liu](http://www.nlpr.ia.ac.cn/iva/liujing/index.html), [Haijie Tian](https://github.com/tianhaijie), [Yong Li](http://www.foreverlee.net/), Yongjun Bao, Zhiwei Fang,and Hanqing Lu

## Introduction

We propose a Dual Attention Network (DANet) to adaptively integrate local features with their global dependencies based on the self-attention mechanism. And we achieve new state-of-the-art segmentation performance on three challenging scene segmentation datasets, i.e., Cityscapes, PASCAL Context and COCO Stuff-10k dataset.

![image](img/overview.png)

## Cityscapes testing set result

We train our DANet-101 with only fine annotated data and submit our test results to the official evaluation server.

![image](img/tab3.png)

## Updates

<font color="#dd0000">**2020/9**：</font>**Renew the code**, which supports **Pytorch 1.4.0** or later!

2020/8：The new TNNLS version DRANet achieves [**82.9%**](https://www.cityscapes-dataset.com/method-details/?submissionID=4792) on Cityscapes test set (submit the result on August, 2019), which is a new state-of-the-arts performance with only using fine annotated dataset and Resnet-101. The code will be released in [DRANet](<https://github.com/junfu1115/DRAN>).

2020/7：DANet is supported on [MMSegmentation](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/danet), in which DANet achieves **80.47%** with single scale testing and **82.02%** with multi-scale testing on Cityscapes val set.

2018/9：DANet released. The trained model with ResNet101 achieves 81.5% on Cityscapes test set.

## Usage

1. Install pytorch 

   - The code is tested on python3.6 and torch 1.4.0.
   - The code is modified from [PyTorch-Encoding](https://github.com/zhanghang1989/PyTorch-Encoding). 

2. Clone the resposity

   ```shell
   git clone https://github.com/junfu1115/DANet.git 
   cd DANet 
   python setup.py install
   ```

3. Dataset
   - Download the [Cityscapes](https://www.cityscapes-dataset.com/) dataset and convert the dataset to [19 categories](https://github.com/mcordts/cityscapesScripts/blob/master/cityscapesscripts/helpers/labels.py). 
   - Please put dataset in folder `./datasets`

4. Evaluation for DANet

   - Download trained model [DANet101](https://drive.google.com/open?id=1XmpFEF-tbPH0Rmv4eKRxYJngr3pTbj6p) and put it in folder `./experiments/segmentation/models/`

   - `cd ./experiments/segmentation/`

   - For single scale testing, please run:

   - ```shell
     CUDA_VISIBLE_DEVICES=0,1,2,3 python test.py --dataset citys --model danet --backbone resnet101 --resume  models/DANet101.pth.tar --eval --base-size 2048 --crop-size 768 --workers 1 --multi-grid --multi-dilation 4 8 16 --os 8 --aux --no-deepstem
     ```

   - Evaluation Result

     The expected scores will show as follows: DANet101 on cityscapes val set (mIoU/pAcc): **79.93/95.97**(ss) 

5. Evaluation for DRANet

   - Download trained model [DRANet101](https://drive.google.com/file/d/1xCl2N0b0rVFH4y30HCGfy7RY3-ars7Ce/view?usp=sharing) and put it in folder `./experiments/segmentation/models/`

   - Evaluation code is in folder `./experiments/segmentation/`

   - `cd ./experiments/segmentation/`

   - For single scale testing, please run:

   - ```shell
     CUDA_VISIBLE_DEVICES=0,1,2,3 python test.py --dataset citys --model dran --backbone resnet101 --resume  models/dran101.pth.tar --eval --base-size 2048 --crop-size 768 --workers 1 --multi-grid --multi-dilation 4 8 16 --os 8 --aux
     ```

   - Evaluation Result

     The expected scores will show as follows: DRANet101 on cityscapes val set (mIoU/pAcc): **81.63/96.62** (ss) 

## Citation

if you find DANet and DRANet useful in your research, please consider citing:

```
@article{fu2020scene,
  title={Scene Segmentation With Dual Relation-Aware Attention Network},
  author={Fu, Jun and Liu, Jing and Jiang, Jie and Li, Yong and Bao, Yongjun and Lu, Hanqing},
  journal={IEEE Transactions on Neural Networks and Learning Systems},
  year={2020},
  publisher={IEEE}
}
```

```
@inproceedings{fu2019dual,
  title={Dual attention network for scene segmentation},
  author={Fu, Jun and Liu, Jing and Tian, Haijie and Li, Yong and Bao, Yongjun and Fang, Zhiwei and Lu, Hanqing},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  pages={3146--3154},
  year={2019}
}
```



## Acknowledgement

Thanks [PyTorch-Encoding](https://github.com/zhanghang1989/PyTorch-Encoding), especially the Synchronized BN!