The code implements the "Locality-Aware Hyperspectral Classification (BMVC2023)"

$python main.py --dataset='Indian' --epoches=300 --patches=7 --band_patches=1 --mode='CAF' --weight_decay=5e-3 --flag='train' --output_dir='./logs/' --batch_size=32 --align='align' --spatial_attn

$python visualization.py

For a detailed experimental setup and dataset information, please refer to our supplementary materials.

The code is built upon SpectralFormer and MAEST, thanks to their great work! If you find it is useful for your research, please kindly cite the following papers:

• Zhou et al. (2023) - Locality-Aware Hyperspectral Classification
• Hong et al. (2021). - SpectralFormer: Rethinking Hyperspectral Image Classification With Transformers
• Damian et al. (2022) - Masked Auto-Encoding Spectral–Spatial Transformer for Hyperspectral Image Classification