Hello, I want to know is there some data process before run data_find.py?

[JingqiNiu]

When I see your paper, I find it in the paper, a IDirD image resize to 768×768 , splitinto 3*3 patch and covert to gray space. I don't find any process in data_find.py. So should I process IDRID frist into that format?

[WentianZhang-ML]

Sure. You need to resize a IDRiD image to 768×768 pixels and convert it to gray. Then use data_find.py to process data and split it into patches.

[JingqiNiu]

So the data_find.py will split it into patches？

[JingqiNiu]

And can you prove the train val split for ADAM and IDIRD?

[WentianZhang-ML]

We provide an example of split operation in function 'hog_split' in data_find.py. And the input of model is in form of patches.

[WentianZhang-ML]

We followed the original protocol of two datasets and you can see the detailed information in their papers.

[JingqiNiu]

Thanks a lot

[WentianZhang-ML]

If you find it is useful, hopefully give a star to this repository