OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

BACKGROUNDMATTINGV2

BACKGROUNDMATTINGV2 is an open-source computer-vision model and repository for high-resolution foreground matting. It uses an additional captured image of the background to separate subjects from scenes in images and video.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

BACKGROUNDMATTINGV2 — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you can provide a separate captured background and need to extract a subject from images or video. It is also suitable for exploring webcam matting or model training; production video pipelines require additional engineering beyond the supplied inference script.

Who it is for

  • Developers building image or video compositing workflows that can capture a clean background.
  • Researchers and practitioners exploring foreground matting or training the model.
  • Linux users interested in the separately documented webcam-to-virtual-camera demo.

Documented uses

  • Generate a foreground matte from a directory of images.
  • Process video with the model for matting.
  • Experiment with webcam matting through the repository demo.
  • Train the network using a configured dataset path.
Before you choose

Requirements & limits

The repository is public and the project is licensed under MIT.

  • The model requires an additional captured background image.
  • Reported performance is 4K at 30 fps and HD at 60 fps on an Nvidia RTX 2080 Ti; these figures are hardware-specific.
  • The video inference script is not intended as a real-time production pipeline and does not use hardware-accelerated or parallel video encoding and decoding.
  • The virtual-camera webcam plugin is a separate community project and is documented for Linux.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare inputs for BackgroundMattingV2

Capture the additional background image needed for background matting.

Before you start

  • A scene for which a background image can be captured.

Steps

  1. Capture an image of the background without the foreground subject.

  2. Use the captured background image as the additional background input when preparing to use the model.

Expected outcome

A captured background image is ready as an input for the model.

Guide limitations

  • The model requires an additional captured background image.
  • The repository lists image, video, and webcam inference scripts, but their detailed usage instructions are in the model usage documentation.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
PeterL1n/BackgroundMattingV2 · Official homepage title · retained in its source language
How we review sourcesSuggest a correction