Sammie-Roto is designed to be a user-friendly tool for AI assisted rotoscoping (masking) of video clips. It serves as a free alternative to commercial solutions such as Adobe’s Roto Brush or DaVinci Resolve Magic Mask. It is generally less accurate than manual rotoscoping, but can usually give a pretty good result with little effort.
You may also be interested in another similar tool that I created, Cutie-Roto. While Cutie-Roto and Sammie-Roto both serve a similar purpose, their internal workings are different, so each one might work better in certain situations.
- Includes 3 segmentation models: Sam 2.1 Large, Sam 2.1 Base+, and EfficientTAM_s_512x512.
- Easy to use interface allows seeking through the video and clicking to add points.
- Multi-object support.
- Various mask postprocessing options, including edge-smoothing.
- Multiple export options: Black and White Matte, Alpha channel, and Greenscreen.
- It is recommended to have at least 10GB of disk space on the drive that you install the application to.
- An NVIDIA GPU with at least 6GB VRAM is highly recommended. The application can run on CPU, but will be extremely slow.
- Download latest version from releases
- Extract the zip archive.
- Run 'install_dependencies.bat' and follow the prompt.
- Run 'run_sammie.bat' to launch the software.
I can only test on Windows, so please let me know if there are any issues with this running on Linux or Mac.
If you want to install on Windows, see the section above!
git clone https://github.com/Zarxrax/Sammie-Roto.git
cd Sammie-Roto
pip install wheel
pip install -r requirements.txt
python sammie\download_models.py
python app.py