liuzuo-byte / DMuCA

A Dual Multi-head Contextual Attention Network for Hyperspectral Image Classification

About this paper

A Dual Multi-head Contextual Attention Network for Hyperspectral Image Classification

Publish www.mdpi.com/journal/remotesensing

The detailed of DMuCA can be seen in the A Dual Multi-head Contextual Attention Network for Hyperspectral Image Classification.

If our code is helpful to you, please cite

Liang M, He Q, Yu X, Wang H, Meng Z, Jiao L. A Dual Multi-Head Contextual Attention Network for Hyperspectral Image Classification. Remote Sensing. 2022; 14(13):3091.

Network:

Figure 1. The structure of the DBDA network.

Figure 2. The structure of the SaMCA block.

Figure 3. The structure of the SeMCA block.

Dataset

You can specify the dataset to be loaded by using the --dataset option. Dataset to train. Available:

1. PaviaU
2. IdianPines
3. Houston

If you want to train on the dataset, please use the following command in the terminal.

python main.py --dataset IndianPines --lr 0.005 --epoch 100  --run 1 --training_percentage 0.10 --class_balancing --patch_size 11