# PeterL1n/RobustVideoMatting

Robust Video Matting in PyTorch, TensorFlow, TensorFlow.js, ONNX, CoreML!

Repository: https://github.com/PeterL1n/RobustVideoMatting
Canonical: https://ross.abutalabs.com/products/robustvideomatting
Homepage: https://peterl1n.github.io/RobustVideoMatting/
Language: Python
License: GPL-3.0
License Family: copyleft
Topics: computer-vision, machine-learning, ai, matting, deep-learning
Last push: 2024-04-02T16:26:48+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1829, "days_push": 883, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 9500, forks 1194 (observed 2026-08-28T04:10:31.501023+00:00)

## What it is
Robust Video Matting (RVM) is a deep learning model and library for real-time human video matting, using a recurrent neural network with temporal memory instead of frame-by-frame processing. It ships pretrained weights for PyTorch, TensorFlow, TensorFlow.js, ONNX, and CoreML, achieving 4K 76FPS on a GTX 1080 Ti.

## Use cases
- remove background from videos without a green screen
- real-time human video matting for webcam effects
- extract alpha mattes of people from video footage
- run portrait segmentation in the browser with TensorFlow.js
- convert videos with transparent human foreground for compositing
- deploy video matting on mobile with CoreML or ONNX

## When to choose
- you need real-time, high-resolution human matting on video streams
- you want pretrained models across many inference frameworks (PyTorch, TF, ONNX, CoreML, TF.js)
- your target is people/humans specifically and you have no trimap or extra inputs

## When to avoid
- you need matting of general objects or animals, not humans
- you need image (single-frame) matting with fine hair detail and no temporal context
- your project cannot comply with the GPL-3.0 license

## Facets
- artifact type: library
- maturity: stable
- function: computer-vision, image-processing, machine-learning, deep-learning
- domain: computer-vision, machine-learning, image-processing
- platform: python, cross-platform, browser, wasm
- tags: video-matting, background-removal, green-screen, human-segmentation, real-time-inference, pretrained-models, onnx, tensorflow-js, coreml, video, gpu

## Member repositories
- PeterL1n/RobustVideoMatting (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:31.501023+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:22:01.601239+00:00, confidence not recorded.
  - readme: https://github.com/PeterL1n/RobustVideoMatting (fetched 2026-08-28T04:10:31.501023+00:00, sha 5ba3e94a3218)
  - homepage: https://peterl1n.github.io/RobustVideoMatting/ (fetched 2026-08-29T08:21:54.192864+00:00, sha 44136fa355b3)
- Data as of 2026-08-30T08:39:29.467469+00:00.