Example 2 - BiRefNet
Updated: 29 Jun 2026
Example 2 - BiRefNet
Please note that we provide an integration to load ONNX AI models in Notch. However, we do not provide support for creating, training, modifying, or troubleshooting ONNX models. We do not guarantee that all ONNX models will be compatible with Notch, specific GPU architectures or driver versions. The AI model space is highly dynamic and you should undertake significant testing before attempting to deploy a specific model in your production environment. See Overview and Setup of the manuals for a guide to the setup of, and requirements of, ONNX models.
There is also complete integrations of AI models already inside of Notch, for example in the Yolo Tracker, AI Hand Tracker and AI Face Tracker nodes. There is information on using those nodes in their respective sections. But this guide is focussed on using other models.
Model #
AI Model BiRefNetExternal
BiRefNet is a high-precision dichotomous image segmentation model. The standard model takes in an RGB image resized to 1024 x 1024 pixels, and performs foreground/background segmentation, outputting a binary alpha matte. We recommend using the 512 x 512 version.
Download 512x512 ONNX Model
from Hugging Face
Disclaimer
As explained in our website terms (which apply to your use of the manual) we do not verify or review the content of external sites (such as those pages and/or files linked below). If you follow these links, you will leave our website, and the terms of the destination website may apply. We include these links for information only, and do not endorse or recommend the linked content, models, or software, nor make any representation or warranty of any kind regarding their accuracy, reliability, security, suitability, or otherwise. You acknowledge that you rely on, use or download linked materials entirely at your own risk.
BiRefNet is a high-precision dichotomous image segmentation model created by Peng Zheng, Dehong Gao, Deng-Ping Fan, Li Liu, Jorma Laaksonen, Wanli Ouyang, and Nicu Sebe, and is available under the MIT License.
The standard model takes in an RGB image resized to 1024 x 1024 pixels, and performs foreground/background segmentation, outputting a binary alpha matte. But we will be using the 512 x 512 version as an ONNX model which can be downloaded from Here
Example Project #
Download the above model and the below example project.
Download Example ProjectImport the model into Notch and then set it as the model in the AI Model Post-FX node.
Model Usage #
This model should be used with the AI Model Post-FX node and used with a video source such as Video Loader.
