Skip to content

lironui/MAResU-Net

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Multi-stage Attention ResU-Net for Semantic Segmentation of Fine-Resolution Remote Sensing Images

Welcome to my HomePage

In this repository, we re-factor the skip connections in the raw U-Net and design a Multi-stage Attention ResU-Net (MAResU-Net) for semantic segmentation from fine-resolution remote sensing images.

The detailed results can be seen in the Multi-stage Attention ResU-Net for Semantic Segmentation of Fine-Resolution Remote Sensing Images.

The training and testing code can refer to GeoSeg.

The related repositories include:

If our code is helpful to you, please cite:

R. Li, S. Zheng, C. Duan, J. Su and C. Zhang. "Multistage Attention ResU-Net for Semantic Segmentation of Fine-Resolution Remote Sensing Images." in IEEE Geoscience and Remote Sensing Letters, doi: 10.1109/LGRS.2021.3063381.

R. Li, C. Duan, S. Zheng, C. Zhang and P. M. Atkinson, "MACU-Net for Semantic Segmentation of Fine-Resolution Remotely Sensed Images," in IEEE Geoscience and Remote Sensing Letters, doi: 10.1109/LGRS.2021.3052886.

Acknowlegement:

Thanks very much for the sincere help from Jianlin Su as well as his blog 线性Attention的探索:Attention必须有个Softmax吗?

Requirements:

numpy >= 1.16.5
PyTorch >= 1.3.1
sklearn >= 0.20.4
tqdm >= 4.46.1
imageio >= 2.8.0

Network:

network
Fig. 1. The structure of (a) the proposed MAResU-Net and (b) the attention block.

Result:

Result
Fig. 2. Visualization of results on the Vaihingen.

Complexity:

Complexity
Fig. 3. The (a) computation requirement and (b) memory requirement of the raw dot-product attention mechanism and the proposed linear attention mechanism under different input sizes. Please notice that the figure is in log scale.