Blind gaussian denoising

[Shakarim94]

Hello! Thank you for this amazing code!

I want to ask about the DnCNN-B model. How one would implement it using this code?

In the paper they say that the patch size is set to 50x50 and the number of layers increased to 20. They also nearly doubled the number of layers. What about the noise level? Do we just pick random sigma between [0,50] at each batch?

[SaoYan]

My PyTorch implementation maybe helpful.

https://github.com/SaoYan/DnCNN-PyTorch

Please refer to line 70~75 in train.py

[qwbmiao]

Hello!Your code looks so cool!

I want to ask you that how I do to implement blind guassian noise by using the code.I try to change the sigma by "feed" like learning rate(lr) in the code,but failed.

self.X = self.Y_ + tf.random_normal(shape=tf.shape(self.Y_), stddev=self.sigma)

So can you tell me how to implement it?Thank you very much

[clausmichele]

In my fork I implemented blind denoising with 0-50 levels for RGB images https://github.com/clausmichele/DnCNN-tensorflow

[liting11]

hi,your code is very cool that is modified.But I went wrong when I work it.I run the main program directly.I don't know that reason is what.can you tell me why?could you tell your email?i want to learn for you.thank you so much

[clausmichele]

Can you explain more in detail what kind of problem you have? It is impossible to say what went wrong without any details of where you are running the software.

[liting11]

I am very happy to receive your reply.thank you very much.you are very good man .I can't run the program you modified。When I run main.py，there was an error about 'TypeError: 'range' object does not support item assignment',I think there are questions about my opencv.I installed OpenCV on the software of spyder, but I couldn't use it.I don't why？You can help me analyze it？Thanks for dear.

[clausmichele]

Please post the full error you get, so I can see the line number and try to understand what's the problem. By the way, I am using OpenCV 3.4 compiled from source, because the python version available is older, but this should not be the problem.

[liting11]

This is result about when runned 'main.py' directly.
runfile('F:/dncnn/DnCNN-tensorflow-master/DnCNN-tensorflow-master/main.py', wdir='F:/dncnn/DnCNN-tensorflow-master/DnCNN-tensorflow-master')
GPU

Traceback (most recent call last):

File "", line 1, in
runfile('F:/dncnn/DnCNN-tensorflow-master/DnCNN-tensorflow-master/main.py', wdir='F:/dncnn/DnCNN-tensorflow-master/DnCNN-tensorflow-master')

File "D:\Anaconda3\lib\site-packages\spyder\utils\site\sitecustomize.py", line 866, in runfile
execfile(filename, namespace)

File "D:\Anaconda3\lib\site-packages\spyder\utils\site\sitecustomize.py", line 102, in execfile
exec(compile(f.read(), filename, 'exec'), namespace)

File "F:/dncnn/DnCNN-tensorflow-master/DnCNN-tensorflow-master/main.py", line 84, in
tf.app.run()

File "D:\Anaconda3\lib\site-packages\tensorflow\python\platform\app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))

File "F:/dncnn/DnCNN-tensorflow-master/DnCNN-tensorflow-master/main.py", line 56, in main
model = denoiser(sess)

File "F:\dncnn\DnCNN-tensorflow-master\DnCNN-tensorflow-master\model.py", line 41, in init
self.dataset = dataset(sess)

File "F:\dncnn\DnCNN-tensorflow-master\DnCNN-tensorflow-master\model.py", line 204, in init
random.shuffle(ind)

File "D:\Anaconda3\lib\random.py", line 272, in shuffle
x[i], x[j] = x[j], x[i]

TypeError: 'range' object does not support item assignment

[clausmichele]

Running main.py only starts the training, is what you want? Do you have the images in the correct folders? Try main.py --phase=test with some test images in the right folders first.

[liting11]

yes,I want to run your code to strart the training.I haved the images in the correct folders.Did your code work successful？Result?

[clausmichele]

Check that ind isn't empty or None and try without random.shuffle(ind)

[Pichairen]

@clausmichele Hi,When I run main.py --phase=test,there is an error happen, AssertionError: No testing data! whether I need download other testing data?

[Shakarim94]

Hello guys, I recently implemented CDnCNN-B (blind color denoising) based on this repository