⭐ ⭐ ⭐ ⭐ are welcome. New tools will be added and existing will be improved with time.
Updates:
[January 9] Added image inpainting. (Existing users should be able to update.)
[November 28] Added interpolate-frames.
[October 31] Use super-resolution as a filter for medium/large images.
[October 17] Added image enlightening.
[September 27] Added Force CPU use button and minor bug fixes.
[August 28] Added deep learning based dehazing and denoising.
[August 25] Simplified installation and updating method.
[August 2] Added deep matting and k-means.
[July 17] MonoDepth and Colorization models have been updated.
[1] Install GIMP 2.10.
[2] Clone this repository: git clone https://github.com/kritiksoman/GIMP-ML.git
[3] Open terminal, go to GIMP-ML/gimp-plugins and run :
bash installGimpML.sh
[4] Open GIMP and go to Preferences -> Folders -> Plug-ins, add the folder gimp-plugins and restart GIMP.
[5] Go to Layer->GIMP-ML->update, click on ok with "update weights" set to yes and restart GIMP. (Weights ~ 1.5GB will be downloaded)
Manual install description if above is not working: Link
[1] Go to Layer->GIMP-ML->update, click on ok with "update weights" set to NO and restart GIMP.
[2] Go to Layer->GIMP-ML->update, click on ok with "update weights" set to YES and restart GIMP.
Please cite using the following bibtex entry:
@article{soman2020GIMPML,
title={GIMP-ML: Python Plugins for using Computer Vision Models in GIMP},
author={Soman, Kritik},
journal={arXiv preprint arXiv:2004.13060},
year={2020}
}
| Name | License | Dataset |
|---|---|---|
| facegen | CC BY-NC-SA 4.0 | CelebAMask-HQ |
| deblur | BSD 3-clause | GoPro |
| faceparse | MIT | CelebAMask-HQ |
| deepcolor | MIT | ImageNet |
| monodepth | MIT | Multiple |
| super-resolution | MIT | ImageNet |
| deepmatting | Non-commercial purposes | Adobe Deep Image Matting |
| semantic-segmentation | MIT | COCO |
| kmeans | BSD | - |
| deep-dehazing | MIT | Custom |
| deep-denoising | GPL3 | BSD68 |
| enlighten | BSD | Custom |
| interpolate-frames | MIT | HD |
| inpainting | CC BY-NC-SA 4.0 | DIV2K |