IOPaint is a free and open-source image restoration and enhancement tool supported by state-of-the-art AI models, making image editing simpler and more efficient. It features functions such as Erase, Replace Object, Draw Text, and Out-painting, allowing users to remove unwanted objects, defects, watermarks, or people from images, or replace anything in the picture.
Open Source Repository: https://github.com/Sanster/IOPaint
Demo:
https://huggingface.co/spaces/Sanster/iopaint-lama
https://colab.research.google.com/drive/1TKVlDZiE3MIZnAUMpv2t_S4hLr6TUY1d?usp=sharing
This project is developed by Sanster and has already received 17.6k stars, indicating its popularity as an open-source project. IOPaint is like a powerful AI toolbox that meets the Photoshop needs of everyone from ordinary users to professional designers.
What's so amazing about it?
- Completely free and open-source, supports CPU, GPU, and Apple Silicon, can be deployed on your own computer
- Convenient for novice users, provides a one-click installation program for Windows.
- Erase and repair content in images, can switch between multiple AI models to compare results
- Integrates multiple popular AI models, such as runwayml/stable-diffusion-inpainting and Sanster/AnyText, etc.
- Has a variety of plugins, such as Segment Anything, RemoveBG, Anime Segmentation, RealESRGAN, GFPGAN, etc., for object segmentation, background removal, anime image segmentation, super-resolution, and facial restoration, etc.
- File manager function, allowing users to browse images and save directly to the output directory.
How to use?
If you are using a Windows system, you can directly go to the open-source project homepage to find the download link. Here is the general deployment method:
- Install IOPaint and dependencies:
pip3 install iopaint - Start the web interface of IOPaint:
iopaint start --model=lama --device=cpu --port=8080 - Access http://localhost:8080 in your browser to start using IOPaint.