ScribbleSeg: Scribble-based Interactive Image Segmentation

TL;DR

ScribbleSeg introduces a new framework for scribble-based interactive image segmentation, with a standardized protocol, a novel training and evaluation strategy, and demonstrates superior performance over click-based methods.

Contribution

The paper formulates a standard protocol for scribble-based segmentation, proposes a new framework with specialized modules, and establishes a challenging benchmark for evaluation.

Findings

ScribbleSeg outperforms previous click-based segmentation methods.

A deterministic scribble generator improves evaluation consistency.

The framework demonstrates robustness across diverse images.

Abstract

Interactive segmentation enables users to extract masks by providing simple annotations to indicate the target, such as boxes, clicks, or scribbles. Among these interaction formats, scribbles are the most flexible as they can be of arbitrary shapes and sizes. This enables scribbles to provide more indications of the target object. However, previous works mainly focus on click-based configuration, and the scribble-based setting is rarely explored. In this work, we attempt to formulate a standard protocol for scribble-based interactive segmentation. Basically, we design diversified strategies to simulate scribbles for training, propose a deterministic scribble generator for evaluation, and construct a challenging benchmark. Besides, we build a strong framework ScribbleSeg, consisting of a Prototype Adaption Module(PAM) and a Corrective Refine Module (CRM), for the task. Extensive experiments show that ScribbleSeg performs notably better than previous click-based methods. We hope this could serve as a more powerful and general solution for interactive segmentation. Our code will be made available.