Added support for histogram matching #129
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Added the option of using histogram matching as a type of style loss. Can be used in conjunction with gram/mean based loss to improve detail of style texture, or can be used on its own to create stylizations that strongly preserve the content image features.
This module is used to match the histogram of the content image features to the histogram of the style image features. Currently does not support GPU acceleration, however it can help achieve more detailed style images when used in conjunction with gram based style loss, and can style images with minimal impact to content features when used as the sole stylization method.
updated slow neural style to have the option of using a histogram loss function.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I've added support for the histogram matching algorithm as described in the following paper
https://arxiv.org/abs/1701.08893
When used in conjunction with gram or mean based style transfer, it helps stabilize style textures within the generated image. When used on its own, i've found that it preserves content image features much better than gram-based style transfer does, while still reproducing a style transfer of high quality.
Currently only implemented as a CPU based method, as creating an openCL or CUDA kernel for the histogram matching algorithm is non-trivial, and will probably take me a while to complete.