madonokouki / block-wise-scrambled-image-recognition Goto Github PK

Thanks alot for sharing your codes. it's a meaningful work.

I have a question about the output of Adaptation Network. For cifar10 dataset, the output size of the adaptive network is 32 * 32 * 16. Why did you select 16 for the channel number? Is choosing a number smaller or larger than 16 OK? How do different values affect the results？

：） ：）

thank you for the code!
I need key files to run proposed_{}_ELE.py
But I cannot find key files or code for generating key files.
How can I make key files?
is there any rule for making key files?

Dear author，
The network is a little large. Whther it is ok to compare with you in smaller network, such as resnet.
Your method and our method is the same as resnet-18.

Traceback (most recent call last):
File "E:\Encode\Anaconda\envs\encrypted_image\lib\site-packages\PIL\Image.py", line 2645, in fromarray
mode, rawmode = _fromarray_typemap[typekey]
KeyError: ((1, 1, 3, 4), '|u1')

im = Image.fromarray(np.uint8(cm.gist_earth(data[0,:,:,:],bytes=True))*255)
File "E:\Encode\Anaconda\envs\encrypted_image\lib\site-packages\PIL\Image.py", line 2647, in fromarray
raise TypeError("Cannot handle this data type")
TypeError: Cannot handle this data type