summitgao / hc_adgan

Code for the paper "Adaptive Dropblock Enhanced GenerativeAdversarial Networks for Hyperspectral Image Classification", IEEE TGRS 2021

This repo holds the codes of our paper:

Adaptive Dropblock Enhanced GenerativeAdversarial Networks for Hyperspectral Image Classification

in IEEE Transactions on Geoscience and Remote Sensing, vol. 59, no. 6, pp. 5040-5053, June 2021, doi: 10.1109/TGRS.2020.3015843.

The demo has not been well organized. Please contact me if you meet any problems.

Please cite our paper if you use our codes. Thanks!

If you have any queries, please do not hesitate to contact me ( gaofeng AT ouc.edu.cn ).

More codes can be obtained from http://feng-gao.cn

Requirements: Python >= 3.6, PyTorch and torchvision

The PaviaU.mat and PaviaU_gt.mat stands for the Pavia University dataset and it's corresponding labels respectively, and the PCU.mat is the result of the Pavia University dataset after PCA.

hc_adgan's Issues

Code

Hello:
Can you provide the final version of the methods? I will used for contrast experiment. The results on Indian pines which I get is similar to your paper, but poor performance on Pavia University.

Thank you

请问您的代码中是否需要将输入数据全部归一化到[-0.5, 0.5]的范围呢