Training dataset

[Yuannianzeng]

Could you provide training datasets？

[fkmajiji]

Okay, I will upload the dataset to the cloud disk in a few days.

[Yuannianzeng]

Thank you！

[fkmajiji]

Now you can download the dataset from the website on Readme.

[Yuannianzeng]

How is the data set generated, and why the following error will be reported when training with other .tif images:
UnboundLocalError: local variable 'd' referenced before assignment

[fkmajiji]

I used the libtiff package in python to generate the data.
The saving code:
from libtiff import TIFFfile, TIFFimage
to create a tiff structure from image data
the size of t is CHW
tiff = TIFFimage(t.astype(np.uint8), description='')
to write tiff structure to file
tiff.write_file((path + 'name.tif'), compression='none')
del tiff # flushes data to disk

Sorry, I don't know about this UnboundLocalError.

[yanshuyue]

@Yuannianzeng Have you solved this error yet? UnboundLocalError: local variable 'd' referenced before assignmen

[Yuannianzeng]

You can see this link ' https://github.com/polwork/Mosaic-Convolution-Attention-Network-for-Demosaicing-Multispectral-Filter-Array-Images ', I run this without error.

[yanshuyue]

@Yuannianzeng Is it convenient to give personal emails? I still report an error when I input a 9-channel image.
arr[i:i + d.nbytes] = d
UnboundLocalError: local variable 'd' referenced before assignment

[yanshuyue]

@fkmajiji I would like to ask you how to get tif format images, I got 9 channels of tif images python does not recognize, you convenient to provide the source code?

[fkmajiji]

This seems to be a problem with the libtiff package. It may be that the version is wrong. In fact, once I configured a new environment, I seem to have encountered this problem, too. Can you send me one your 9 channels of tif image? Let me have a check. My email: 291776435@qq.com