search

zhaoyuzhi / HSGAN

HSGAN: Hyperspectral Reconstruction from RGB Images with Generative Adversarial Network. IEEE TNNLS, 2023
17 stars 1 forks source link

dataset #2

Open zahrabsh74 opened 8 months ago

zahrabsh74 commented 8 months ago

dear @zhaoyuzhi , thanks for your nice work. I have a question about the dataset you have used. are these 5 datasets free and have public access for everyone? I wanted to train your code on the same dataset that you had used, but I couldn't find ARAD,ICV, and NUS dataset in the internet and download them. In addition, as another favor, is it possible to please release the test or inference code on your model?

zhaoyuzhi commented 7 months ago

Datasets:

  1. For ARAD dataset, please refer to: https://openaccess.thecvf.com/content_CVPRW_2020/papers/w31/Arad_NTIRE_2020_Challenge_on_Spectral_Reconstruction_From_an_RGB_Image_CVPRW_2020_paper.pdf
  2. For ICVL dataset, it can be downloaded here: https://icvl.cs.bgu.ac.il/hyperspectral-imaging/
  3. For NUS dataset, it can be downloaded here: https://cvil.eecs.yorku.ca/projects/public_html/spectral_reconstruction/
zhaoyuzhi commented 7 months ago

We use the same evaluation code in: https://github.com/boazarad/NTIRE2020_spectral/blob/master/EvalMetrics.py