# Double-Branch-Dual-Attention-Mechanism-Network

**Repository Path**: mantte6199/Double-Branch-Dual-Attention-Mechanism-Network

## Basic Information

- **Project Name**: Double-Branch-Dual-Attention-Mechanism-Network
- **Description**: pyTorch:高光谱图像分类
该存储库基于PyTorch和sklearn实现了6个用于高光谱图像分类的框架。
详细结果可见于 基于的高光谱图像分类 双分支双注意机制网络 。

如果您需要任何其他信息，请随时与我联系： lironui@whu.edu.cn 。

我们的一些代码引用了项目

    双重注意网络
    遥感影像分类
    用于高光谱图像分类的快速密集光谱空间卷积网络框架

- **Primary Language**: Python
- **License**: AGPL-3.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 9
- **Forks**: 3
- **Created**: 2020-10-01
- **Last Updated**: 2025-04-29

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Hyperspectral Image Classification

This repository implementates 6 frameworks for hyperspectral image classification based on PyTorch and sklearn.

The detailed results can be seen in the [Classification of Hyperspectral Image Based on 
Double-Branch Dual-Attention Mechanism Network](https://www.mdpi.com/2072-4292/12/3/582).

Feel free to contact me if you need any further information: lironui@whu.edu.cn.

Some of our code references the projects
* [Dual-Attention-Network](https://github.com/SH8899/Dual-Attention-Network.git)
* [Remote sensing image classification](https://github.com/stop68/Remote-Sensing-Image-Classification.git)
* [A Fast Dense Spectral-Spatial Convolution Network Framework for Hyperspectral Images Classification](https://github.com/shuguang-52/FDSSC.git) 

If our code is helpful to you, please cite
`Li R, Zheng S, Duan C, et al. Classification of Hyperspectral Image Based on Double-Branch Dual-Attention Mechanism Network[J]. Remote Sensing, 2020, 12(3): 582.`


Requirements：
------- 
```
numpy >= 1.16.5
PyTorch >= 1.3.1
sklearn >= 0.20.4
```

Datasets:
------- 
You can download the hyperspectral datasets in mat format at: http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes, and move the files to `./datasets` folder.

Usage:
------- 
1. Set the percentage of training and validation samples by the `load_dataset` function in the file `./global_module/generate_pic.py`.
2. Taking the DBDA framework as an example, run `./DBDA/main.py` and type the name of dataset. 
3. The classfication maps are obtained in `./DBDA/classification_maps` folder, and accuracy result is generated in `./DBDA/records` folder.

Network:
------- 
* [DBDA](https://www.mdpi.com/2072-4292/12/3/582)
* [DBMA](https://www.mdpi.com/2072-4292/11/11/1307/xml)
* [FDSSC](https://www.mdpi.com/2072-4292/10/7/1068/htm)
* [SSRN](https://ieeexplore.ieee.org/document/8061020)
* [CDCNN](https://ieeexplore.ieee.org/document/7973178)
* [SVM](https://ieeexplore.ieee.org/document/1323134/)

![network](https://github.com/lironui/Double-Branch-Dual-Attention-Mechanism-Network/blob/master/figures/Figure%206.%20The%20structure%20of%20the%20DBDA%20network.png)  
Figure 1. The structure of the DBDA network. The upper Spectral Branch composed of the dense 
spectral block and channel attention block is designed to capture spectral features. The lower Spatial 
Branch constituted by dense spatial block, and spatial attention block is designed to exploit spatial 
features. 

Results:
------- 
![IP](https://github.com/lironui/Double-Branch-Dual-Attention-Mechanism-Network/blob/master/figures/Figure%209.%20Classi%EF%AC%81cation%20maps%20for%20the%20IP%20dataset%20using%203%25%20training%20samples.png)
Figure 2. Classification maps for the IP dataset using 3% training samples. (a) False-color image. (b) 
Ground-truth (GT). (c)–(h) The classification maps with disparate algorithms. 
![UP](https://github.com/lironui/Double-Branch-Dual-Attention-Mechanism-Network/blob/master/figures/Figure%2010.%20Classi%EF%AC%81cation%20maps%20for%20the%20UP%20dataset%20using%200.5%25%20training%20samples.png)
Figure 3. Classification maps for the UP dataset using 0.5% training samples. (a) False-color image. 
(b) Ground-truth (GT). (c)–(h) The classification maps with disparate algorithms. 
![SV](https://github.com/lironui/Double-Branch-Dual-Attention-Mechanism-Network/blob/master/figures/Figure%2011.%20Classi%EF%AC%81cation%20maps%20for%20the%20SV%20dataset%20using%200.5%25%20training%20samples.png)
Figure 4. Classification maps for the UP dataset using 0.5% training samples. (a) False-color image. 
(b) Ground-truth (GT). (c)–(h) The classification maps with disparate algorithms. 
![BS](https://github.com/lironui/Double-Branch-Dual-Attention-Mechanism-Network/blob/master/figures/Figure%2012.%20Classi%EF%AC%81cation%20maps%20for%20the%20BS%20dataset%20using%201.2%25%20training%20samples.png)
Figure 5. Classification maps for the BS dataset using 1.2% training samples. (a) False-color image. 
(b) Ground-truth (GT). (c)–(h) The classification maps with disparate algorithms.