Skip to content
/ DCN-T Public

[TIP 2023] The official repo for the paper "DCN-T: Dual Context Network with Transformer for Hyperspectral Image Classification".

Notifications You must be signed in to change notification settings

DotWang/DCN-T

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DCN-T: Dual Context Network with Transformer for Hyperspectral Image Classification (TIP 2023)

Di Wang, Jing Zhang, Bo Du, Liangpei Zhang and Dacheng Tao

Pytorch implementation of our paper for ImageNet Pretraining and Transformer-based image-level hyperspectral image classification.


Fig.1 - The proposed DCN-T.

Fig.2 - The DCM.

Usage

  1. Install Pytorch 1.9 with Python 3.8.
  2. Clone this repo.
git clone https://github.com/DotWang/DCN-T.git
  1. Prepare the tri-spectral dataset with the notebook
  2. Download ImageNet pretrained model
  3. For implementing the clusttering, install the SSN
cd utils/gensp/src
python setup.py install

Or you can taste the pytorch version realized in the network_local_global.py

  1. Training and Testing

For example, training on the WHU-Hi-LongKou scene with soft voting

CUDA_VISIBLE_DEVICES=0 python -m torch.distributed.launch --nproc_per_node=1 --nnodes 1 \
    --node_rank=0 --master_port=1901 --use_env train_memory.py \
    --dataset 'WHUHi_LongKou_15_100' \
    --backbone 'vgg16' \
    --epochs 30 --lr 1e-3 --groups 128 --eval_interval 1 \
    --batch_size 4 --test_batch_size 1 --workers 2 \
    --ra_head_num 4 --ga_head_num 4 --mode 'soft'
CUDA_VISIBLE_DEVICES=0 python test_gpu.py \
    --dataset 'WHUHi_LongKou_15_100' \
    --backbone 'vgg16' --ra_head_num 4 --ga_head_num 4 \
    --scales 1  --groups 128 \
    --model_path './run/WHUHi_LongKou_15_100/vgg16_128/experiment_0/model_last.pth.tar' \
    --save_folder './run/WHUHi_LongKou_15_100/vgg16_128/experiment_0/'

Citation

@ARTICLE{wang_2023_dcnt,
  author={Wang, Di and Zhang, Jing and Du, Bo and Zhang, Liangpei and Tao, Dacheng},
  journal={IEEE Transactions on Image Processing}, 
  title={DCN-T: Dual Context Network With Transformer for Hyperspectral Image Classification}, 
  year={2023},
  volume={32},
  number={},
  pages={2536-2551},
  doi={10.1109/TIP.2023.3270104}}

Thanks

SSN-PytorchSpixelFCNBIT

Relevant Projects

[1] Pixel and Patch-level Hyperspectral Image Classification
    Adaptive Spectral–Spatial Multiscale Contextual Feature Extraction for Hyperspectral Image Classification, IEEE TGRS, 2020 | Paper | Github
    Di Wang, Bo Du, Liangpei Zhang and Yonghao Xu

[2] Image-level/Patch-free Hyperspectral Image Classification
    Fully Contextual Network for Hyperspectral Scene Parsing, IEEE TGRS, 2021 | Paper | Github
    Di Wang, Bo Du, and Liangpei Zhang

[3] Graph Convolution based Hyperspectral Image Classification
    Spectral-Spatial Global Graph Reasoning for Hyperspectral Image Classification, IEEE TNNLS, 2023 | Paper | Github
    Di Wang, Bo Du, and Liangpei Zhang

[4] Neural Architecture Search for Hyperspectral Image Classification
    HKNAS: Classification of Hyperspectral Imagery Based on Hyper Kernel Neural Architecture Search, IEEE TNNLS, 2023 | Paper | Github
    Di Wang, Bo Du, Liangpei Zhang, and Dacheng Tao