trimap-based FBA matting with strong quality, but not background-image-based or real-time
Official repository for the paper F, B, Alpha Matting
- 502
- Jupyter Notebook
- MIT
image/video matting
Official implementation of Background Matting V2, a real-time high-resolution image/video matting model that requires an additional background image and uses a neural architecture for refinement.
Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.
trimap-based FBA matting with strong quality, but not background-image-based or real-time
Official repository for the paper F, B, Alpha Matting
real-time trimap-free portrait matting without a background image, but not general high-res video matting
A Trimap-Free Portrait Matting Solution in Real Time [AAAI 2022]
end-to-end deep matting without trimap, but not background-based or real-time
[IJCV 2022] Bridging Composite and Real: Towards End-to-end Deep Image Matting
automatic natural image matting without trimap, but not real-time background matting
[IJCAI'21] Deep Automatic Natural Image Matting
semantic-guided human matting with trimap, but not background-image-based video matting
Robust Human Matting via Semantic Guidance, ACCV 2022.
privacy-preserving portrait matting, but not real-time high-res background matting
[ACM MM 2021] Privacy-Preserving Portrait Matting
mask-guided progressive refinement matting, but not background-based or real-time
This repository includes the official project of Mask Guided (MG) Matting, presented in our paper: Mask Guided Matting via Progressive Refinement Network
ViT-based image matting with high quality, but not real-time background matting
[Information Fusion (Vol.103, Mar. '24)] Boosting Image Matting with Pretrained Plain Vision Transformers
An earlier or more limited way to do the job. Still the right call when you need something small, proven, or CPU-only.
older trimap-based Deep Image Matting baseline; choose for simple non-real-time matting, not high-res video
Deep Image Matting
TensorFlow Deep Image Matting implementation; older and unmaintained, not real-time
This is tensorflow implementation for paper "Deep Image Matting"
older U-Net GAN background removal; simpler but not high-resolution or real-time
Background Removal based on U-Net
older semantic human matting method; narrower and stale
Semantic Human Matting
PyTorch Deep Image Matting reimplementation; older and unmaintained
Pytorch implementation of deep image matting
another older PyTorch Deep Image Matting implementation; stale and not real-time
Deep Image Matting implementation in PyTorch
traditional OpenCV sampling-based alpha matting; CPU-only and older than deep models
An OpenCV2 version for image matting algorithm described by papre - Shared Sample for Real-Time Alpha Matting
older guided contextual attention matting; not real-time and unmaintained
Official repository for Natural Image Matting via Guided Contextual Attention
classic closed-form matting algorithm; CPU-based, slow, and previous generation
Python implementation of A. Levin D. Lischinski and Y. Weiss. A Closed Form Solution to Natural Image Matting. IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), June 2006, New York
older IndexNet deep matting; not real-time and unmaintained
(ICCV'19) Indices Matter: Learning to Index for Deep Image Matting
These projects were analysed and named BackgroundMattingV2 among their alternatives. The relationship is not symmetric — how BackgroundMattingV2 rates them is a separate judgement, made when BackgroundMattingV2 is analysed in its own right.
calls BackgroundMattingV2 “Choose this when you have a static background frame and need high-resolution alpha mattes; it solves background removal with a different technique requiring that extra background input.”
The code for our newly accepted paper in Pattern Recognition 2020: "U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection."