What would the weekend on slashCAM be without news about artificial intelligence in image manipulation? Unthinkable! So here is another project to amaze you, which can adapt the image content to the style of the background already during insertion. If you scroll through the linked examples on the GitHub page, you should quickly recognize the potential of this development.
The automatic photomontage is just around the corner
The natural fitting of foreign objects into an image will function automatically in the future. And without special compositing knowledge for everyone. Those who already want to understand the "Deep Painterly Harmonization" themselves cannot rely on Tensorflow, but have to install Torch7 and Matlab/Octave. Or dare to port yourself, but that's probably not quite a slashCAM-theme anymore...
;)
