Re:Deep image prior 'learns' on just one image
by lern_too_spel on Friday December 01, 2017 @12:38AM
(#15822165)
The point is that a fully connected network would not get you the same results. I.e. the convolutional structure allows it to learn interesting things from a single example because by choosing a convolutional architecture, you are implicitly imposing a prior.
Re:Deep image prior 'learns' on just one image
by fjsolwmv on Friday December 01, 2017 @12:16PM
(#15825317)
How is a network structure different from a fully connected network with 0 weights?
Re:Deep image prior 'learns' on just one image
by kalal on Friday December 01, 2017 @03:39AM
(#15822719)
Agreed. The network is trained in a sense "online" on the target image and directly applied to it. I don't see much impact of this work outside image denouncing or tasks already presented in the paper.
> contrary to expectations,
ingore the weasel words
> a great deal of image statistics are captured by the structure of a convolutional image generator rather than by any learned capability.
> To show this, we apply untrained ConvNets to the solution of several such problems. Instead of following the common paradigm of training a ConvNet on a large dataset of example images, we fit a generator network to a single degraded image.
How is that _untrained_, if they are training it on an image? Is a generator network different from a ConvNet?
> This is particularly remarkable because no aspect of the network is learned from data;
but they just said "we fit a generator network to a single degraded image"
As others have commented, they appear to be training on all the parts of a single complex image, and using that training to repair local deviations from the global average. This may be better than classical denoising algorithms because the NN can model the image structure better than other approaches, but this doesn't seem like a novel use of NNs.