U-Net Photo DeNoiser with RawTherapee?

A few weeks ago I viewed “Research at NVIDIA: AI Can Now Fix Your Grainy Photos by Only Looking at Grainy Photos” https://www.youtube.com/watch?v=pp7HdI0-MIo and was impressed with the results.

Downloaded the paper “Noise2Noise: Learning Image Restoration without Clean Data” https://arxiv.org/pdf/1803.04189.pdf which mentioned on p 3 “we switch from RED30 to a shallower
U-Net (Ronneberger et al., 2015) that is roughly 10× faster to train and gives similar results.”

When checked “U-Net: Convolutional Networks for Biomedical Image Segmentation” by Olaf Ronneberger
https://arxiv.org/pdf/1505.04597.pdf I found at his website https://lmb.informatik.uni-freiburg.de/people/ronneber/u-net/ "We provide the u-net for download in the following archive: u-net-release-2015-10-02.tar.gz (185MB). It contains the ready trained network, the source code, the matlab binaries of the modified caffe network, all essential third party libraries, the matlab-interface for overlap-tile segmentation and a greedy tracking algorithm used for our submission for the ISBI cell tracking challenge 2015. Everything is compiled and tested only on Ubuntu Linux 14.04 and Matlab 2014b (x64)

To apply the segmentation and the tracking to the images in “PhC-C2DH-U373/01” simply run the shell script ./segmentAndTrack.sh

The resulting segmentation masks will be written to “PhC-C2DH-U373/01_RES”

If you do not have a CUDA-capable GPU or your GPU is smaller than mine, edit segmentAndTrack.sh accordingly (see there for documentation). If you have any questions, you may contact me at ronneber@informatik.uni-freiburg.de, but be aware that I can not provide any support."

1st I do not code so hope others with this experience can tell me. Can U-Net be added to RawTherapee?

Appreciate all comments!

Thanks

Ken

Technically, yes it could. It could be added to anything. But practically, probably not. You need quite a CUDA-capable card to do this, and RT doesn’t use OpenCL at the moment. You can always process in RT and denoise after the fact, if that suites you.