shuangxu96 / gppnn Goto Github PK

Dear @xsxjtu!

Sorry for interrupting you.
I tried to customize your code to test on Quickbird and Worldview3 satellites.
I generated 8000 pairs of training images, in which the patch size is 128x128 for PAN image and 32x32 for LRMS image.
After training GPPNN using my training set, I tested with the test set, but it works poorly. It is to say, it outperforms other traditional benchmarks but is even worse than early DL work like PanNet.

What parameters can I customize?

Best regards.

您好，我是一个图像融合的初学者，我对您的工作非常感兴趣，我想向您请求一下您在论文的数据集部分（QuickBird、GaoFen以及Landsat），不知道您是否方便。我保证仅限个人学习使用，如果方便的话，我留下我的邮箱：[email protected]，冒昧打扰，十分抱歉。祝好！

Hi @shuangxu96,

Thanks a lot for your code sharing!
I am running your train_GPPNN.py file on my machine and also Google Colab without changing anything (except savemath file directory according to my computer OS's directory character). Unfortunately, I am finding very poor metric results like in figure. I couldn't find why.

My another question is 'Why you use Landsat8_train.h5 file for training data in your code but don't use Landsat8_5.mat file? Are there any differences between them?

If you answer me, I will be very glad.
Thanks in advance.

Hi @xsxjtu

I like your work and thank you for making this repo publically available. However, I'm wondering where to download the datasets used in this paper? If possible, can you please email me the full datasets used in your experiments? My email is [email protected]

Thanks,
Chaminda