Implementation Codes for HyperDID

Introduction

This project provides the implementation of the paper "HyperDID: Hyperspectral Intrinsic Image Decomposition with Deep Feature Embedding". [paper]

Requirements

• torch
• cuda

Project architecture

• demo.py
• data_process.py
• evaluate.py
• model
  • contrastive_learning.py
• data
  • indian
    • IndianPine.mat
    • AVIRIS_colormap.mat
  • pavia
  • houston2013
  • houston2018
• checkpoint
• results

Running

• Training

python demo.py --patches 5

• Testing

python demo.py --patches 5 --flag_test test

Citing this work

If you find this work helpful for your research, please consider citing:

@article{gong2024,
    Author = {Zhiqiang Gong and Xian Zhou and Wen Yao and Xiaohu Zheng and Ping Zhong},
    Title = {HyperDID: Hyperspectral Intrinsic Image Decomposition with Deep Feature Embedding},
    Journal = {IEEE Transactions on Geoscience and Remote Sensing},
    volume = 62,
    Year = {2024}
}