Machine Learning plugins for GIMP 3.

Forked from the original version to improve the user experience in several aspects:

• Added more models.
• Models are run with Python 3.10+.
• Full error text is shown in the GIMP error dailog and in debug console.
• Additional alpha channel handling in some plugins.
• Automatic installation for Windows systems.
• And other smaller improvements.

The plugins have been tested with GIMP 2.99.12 on the following systems:

• Windows 10

1. Install GIMP3.
2. Download this repository.
3. On Windows:
   • Install Python 3.10.
   • Run install.cmd from the unpacked folder.
4. You should now find the GIMP-ML plugins under Layers → GIMP-ML.
5. You can download the weights here, or from the weight links below.

• Source: https://github.com/danielgatis/rembg
• Weights:
  • u2net (download, source): A pre-trained model for general use cases.
  • u2netp (download, source): A lightweight version of u2net model.
  • u2net_human_seg (download, source): A pre-trained model for human segmentation.
  • (unused) u2net_cloth_seg (download, source): A pre-trained model for Cloths Parsing from human portrait. Here clothes are parsed into 3 category: Upper body, Lower body and Full body.
• License: MIT License

• Source: https://github.com/youyuge34/Anime-InPainting
• Weights: Google Drive | Baidu
• License: Creative Commons Attribution-NonCommercial 4.0 International

@inproceedings{nazeri2019edgeconnect,
  title={EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning},
  author={Nazeri, Kamyar and Ng, Eric and Joseph, Tony and Qureshi, Faisal and Ebrahimi, Mehran},
  journal={arXiv preprint},
  year={2019}}

• Source:
  • Demosaics: https://github.com/rekaXua/demosaic_project
  • ESRGAN: https://github.com/xinntao/ESRGAN
• Weights: 4x_FatalPixels
• Licenses:
  • Demosaics: GNU Affero General Public License v3.0
  • ESRGAN: Apache-2.0 license

[Paper]
Xintao Wang, Liangbin Xie, Chao Dong, Ying Shan
Applied Research Center (ARC), Tencent PCG
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences

• Source: https://github.com/a-mos/High_Resolution_Image_Inpainting
• License: Creative Commons Attribution-NonCommercial 4.0 International

@article{Moskalenko_2020,
	doi = {10.51130/graphicon-2020-2-4-18},
	url = {https://doi.org/10.51130%2Fgraphicon-2020-2-4-18},
	year = 2020,
	month = {dec},
	pages = {short18--1--short18--9},
	author = {Andrey Moskalenko and Mikhail Erofeev and Dmitriy Vatolin},
	title = {Deep Two-Stage High-Resolution Image Inpainting},
	journal = {Proceedings of the 30th International Conference on Computer Graphics and Machine Vision ({GraphiCon} 2020). Part 2}} 

• Source: https://github.com/twtygqyy/pytorch-SRResNet
• Torch Hub fork: https://github.com/valgur/pytorch-SRResNet
• License: MIT
• C. Ledig et al., “Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network,” in 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 105–114.

• Source: https://github.com/zeruniverse/neural-colorization
• Torch Hub fork: https://github.com/valgur/neural-colorization
• License:
  • GNU GPL 3.0 for personal or research use
  • Commercial use prohibited
  • Model weights released under CC BY 4.0
• Based on fast-neural-style:
  • https://github.com/jcjohnson/fast-neural-style
  • License:
    • Free for personal or research use
    • For commercial use please contact the authors
  • J. Johnson, A. Alahi, and L. Fei-Fei, “Perceptual Losses for Real-Time Style Transfer and Super-Resolution,” in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9906 LNCS, 2016, pp. 694–711.

• Source: https://github.com/xavysp/DexiNed
• Weights: BIPED
• License: MIT license

@misc{soria2021dexined_ext,
    title={Dense Extreme Inception Network for Edge Detection},
    author={Xavier Soria and Angel Sappa and Patricio Humanante and Arash Arbarinia},
    year={2021},
    eprint={arXiv:2112.02250},
    archivePrefix={arXiv},
    primaryClass={cs.CV}}

• Source: https://github.com/TAMU-VITA/DeblurGANv2
• Torch Hub fork: https://github.com/valgur/DeblurGANv2
• License: BSD 3-clause
• O. Kupyn, T. Martyniuk, J. Wu, and Z. Wang, “DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better,” in 2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 8877–8886.

• Source: https://github.com/nianticlabs/monodepth2
• Torch Hub fork: https://github.com/valgur/monodepth2
• License:
  • See the license file for terms
  • Copyright © Niantic, Inc. 2019. Patent Pending. All rights reserved.
  • Non-commercial use only
• C. Godard, O. Mac Aodha, M. Firman, and G. Brostow, “Digging Into Self-Supervised Monocular Depth Estimation,” in 2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 3827–3837.

• UserUnknownFactor
• Kritik Soman (kritiksoman) – original GIMP-ML implementation

MIT

Please note that additional license terms apply for each individual model. See the references list for details. Many of the models restrict usage to non-commercial or research purposes only.