BEN: Using Confidence-Guided Matting for Dichotomous Image Segmentation

TL;DR

This paper introduces Confidence-Guided Matting (CGM), a novel architecture for dichotomous image segmentation that combines image matting and segmentation to improve boundary accuracy, exemplified by the new BEN model.

Contribution

The paper presents the first CGM model called BEN, integrating matting and segmentation for improved boundary prediction in DIS tasks.

Findings

BEN outperforms state-of-the-art methods on DIS5K dataset

Matting-based refinement significantly enhances segmentation quality

The approach introduces a new paradigm for combining matting and segmentation

Abstract

Current approaches to dichotomous image segmentation (DIS) treat image matting and object segmentation as fundamentally different tasks. As improvements in image segmentation become increasingly challenging to achieve, combining image matting and grayscale segmentation techniques offers promising new directions for architectural innovation. Inspired by the possibility of aligning these two model tasks, we propose a new architectural approach for DIS called Confidence-Guided Matting (CGM). We created the first CGM model called Background Erase Network (BEN). BEN consists of two components: BEN Base for initial segmentation and BEN Refiner for confidence-based refinement. Our approach achieves substantial improvements over current state-of-the-art methods on the DIS5K validation dataset, demonstrating that matting-based refinement can significantly enhance segmentation quality. This work introduces a new paradigm for integrating matting and segmentation techniques, improving fine-grained object boundary prediction in computer vision.