Mask Guided Matting via Progressive Refinement Network
Abstract
- propose Mask Guided (MG) Matting, a robust matting framework that takes a general coarse mask as guidance
- MG Matting leverages a network (PRN) design which encourages the matting model to provide self-guidance to progressively refine the uncertain regions through the decoding process.
- A series of guidance mask perturbation operations are also introduced in the training to further enhance its robustness to external guidance.
- Code and models are available at https://github.com/yucornetto/MGMatting.
Introduction
Image matting is a fundamental computer vision problem which aims to predict an alpha matte to precisely cut out an image region. Most previous matting methods require a well-annotated trimap as an auxiliary guidance input, which explicitly defines the regions of foreground and background as well as the unknown part for the matting methods to solve.
researchers start to study the matting problem in a trimap-free setting.
- One direction is to get rid of any external guidance, and hope that the matting model can capture both semantics and details by end-to end training on large-scale datasets.
- Another line of works investigate alternatives to the trimap guidance, easing the requirement for human input.
Mask Guided (MG) Matting
- In this work, the author introduce a Mask Guided (MG) Matting method which takes a general coarse mask as guidance.
Progressive Refinement Network (PRN) module
- To achieve such robustness to guidance input, the author propose a Progressive Refinement Network (PRN) module, which learns to p