Conventional-to-Transformer-for-Hyperspectral-Image-Classification-Survey-2024

Reference

This code is used for our Survey paper: A Comprehensive Survey for Hyperspectral Image Classification: The Evolution from Conventional to Transformers. The paper is accessible on ArXiv via the following link: "https://arxiv.org/abs/2404.14955v2"

@misc{ahmad2024traditional, title={Traditional to Transformers: A Survey on Current Trends and Future Prospects for Hyperspectral Image Classification}, author={Muhammad Ahmad and Salvatore Distifano and Manuel Mazzara and Adil Mehmood Khan}, year={2024}, eprint={2404.14955}, archivePrefix={arXiv}, primaryClass={cs.CV} }

OR

This code is used for our Survey paper: Traditional to Transformers: A Survey on Current Trends and Future Prospects for Hyperspectral Image Classification. The paper is accessible on ArXiv via the following link: "https://arxiv.org/abs/2404.14945"

@misc{ahmad2024traditional, title={Traditional to Transformers: A Survey on Current Trends and Future Prospects for Hyperspectral Image Classification}, author={Muhammad Ahmad and Salvatore Distifano and Manuel Mazzara and Adil Mehmood Khan}, year={2024}, eprint={2404.14955}, archivePrefix={arXiv}, primaryClass={cs.CV} }

(2+1)D Extreme Xception Net

2D CNN

2D Inception Net

3D Inception Net

2D Xception Net

3D CNN

Hybrid Net

SCSNet

AttentionGraph

Spatial-Spectral Transformer

Requirements

This tool is compatible with Python 2.7 and Python 3.5+ and executed over Colab.

Hyperspectral datasets

Several public hyperspectral datasets are available on the EHU. Users can download those beforehand.

An example dataset folder has the following structure:

Datasets
├── Indian Pines
│   ├── Indian_pines_corrected.mat
│   ├── Indian_pines_gt.mat
├── University of Houston
│   ├── UH.mat
│   └── UG_gt.mat
├── Pavia University
│   ├── PU.mat
│   └── PU_gt.mat
├── Salinas
│   ├── SA.mat
│   └── SA_gt.mat