Our Machine Learning in Linux series focuses on apps that make it easy to experiment with machine learning. All the apps covered in the series can be self-hosted.
RMBG-2-Studio is an enhanced background remove and replace app built around BRIA-RMBG-2.0.
Note that while RMBG-2-Studio is open source software, it’s built on a background removal model (RMBG v2.0) that is released under a Creative Commons license for non-commercial use only.
RMBG v2.0 is designed to effectively separate foreground from background in a range of categories and image types. This model has been trained on a carefully selected dataset, which includes: general stock images, e-commerce, gaming, and advertising content.
Installation
We evaluated RMBG-2-Studio using the Ubuntu 24.10 distro.
In our previous reviews in this series, this section often involves a fairly complex installation process. However, this app is installed with Pinokio, a useful AI utility which makes installation a trivial affair.
In Operation
The app is built with Gradio which provides a friendly web interface.
To get started, drop an image in the input image pane or click the pane to upload an image. Within a few seconds the output is displayed in the adjacent before/after pane. This pane has a slider which visualizes the background removal.
The results are impressive to say the least. We’ve not moved the slider all the way to the right so you can see a small part of the background.
In the above image, the slider is moved all the way to the right.
Results are not perfect though, particularly if the image is complex. For example, the image below features 6 coworkers set in an online group meeting. While the software has made a great job of removing the background it’s also completely erased 2 of the coworkers from the output. Oops!
A Drag-and-Drop gallery resides at the top of the web interface. The gallery lets you view processed images and drag them into composition windows for background replacement and color grading. There are placement control and color grading options, shown below.
The placement controls let you:
- Adjust the size of your image.
- Rotate the image.
- Move the image left/right and/or up/down.
- Mirror the image horizontally.
- Mirror the image vertically.
While not a substitute for a good image editor, we can also apply color grading:
- Adjust the image brightness.
- Adjust image contrast.
- Adjust color intensity.
- Tint color.
- Tint strength.
- Adjust warm/cool color balance.
The software also offers batch processing where you can process multiple images simultaneously. We cab also load images directly from URLs.
Summary
RMBG-2-Studio is a great app to remove backgrounds from images. Installation is trivial, the Gradio interface is easy to use, and results are extremely fast with a good dedicated graphics card.
Built with an enhanced Pinokio interface, the app allows users to efficiently remove backgrounds, replace them with custom compositions, and apply advanced color grading. But this is no replacement for a good image editor.
This tool is useful for content creators, anyone involved in e-commerce business, or creative professionals. JPG, PNG, WEBP, and GIF formats are supported.
RMBG-2-Studio is cross-platform software which runs under Linux, macOS, and Windows. For fast background removal your machine should have a dedicated graphics card.
Website: github.com/pinokiofactory/RMBG-2-Studio
Support:
Developer: RMBG-2-Studio
License: Open source but the background removal model is released under a Creative Commons license for non-commercial use only
For other useful open source apps that use machine learning/deep learning, we’ve compiled this roundup.
RMBG-2-Studio is written in Python. Learn Python with our recommended free books and free tutorials.