ValueError: 'arr' does not have a suitable array shape for any mode. about srcnn-tensorflow

[xiaoshidai]

I have the same problem with you, did you solve this problem?

[xiaoshidai]

I found some people have the same problem and you had update your code, but the problem still exist. Did you test it again?

[xiaoshidai]

There a wrong in the code "util.py"
"""
def imsave(image, path):
return scipy.misc.imsave(path, image)
"""
should be
"""
def imsave(image, path):
return scipy.misc.imsave(image,path)
"""

[tegg89]

@xiaoshidai @vinnitu Sorry for late reply. Because my machine is currently power-off, I cannot check it right now. I will make it right as soon as possible. Thanks

[xiaoshidai]

Did find that the super-resolution image is smaller than the original image, but the code in caffe does not have this problem? Did you have some idea about it?

[tegg89]

@xiaoshidai The size of super-resolved image and that of original image have to be different, because height size and width size of original image were changed due to do proper mod operation.

@vinnitu @xiaoshidai I fixed the problem. Now the result image should be saved.

[HCShi]

@xiaoshidai I also meet this problem ,and I modify the error as you said in util.py , but it doesn't work . So could it be the quality of image or the quantity of the images?

[kjaisinghyadav]

@vinnitu @xiaoshidai What need to be change to fix this probleam??

[ParmpalSinghGill]

check the shape of image before saving