# ScribbleSeg: Scribble-based Interactive Image Segmentation

Source: [ScribbleSeg: Scribble-based Interactive Image Segmentation](<https://www.sota2.com/research/paper/68fa6ed21e209131cea42685>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-28T13:56:58.041Z

Authors: Xi Chen, Yau Shing Jonathan Cheung, Ser-Nam Lim, Hengshuang Zhao

Published: 2023-03-20T17:57:03.000Z

## 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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Road Segmentation on Baseline dataset](<https://www.sota2.com/research/sota/road-segmentation-on-baseline-dataset.md>): Road Segmentation; Baseline dataset; Dice: 76.6.

- [Road Segmentation on WorldRoadSeg 360K](<https://www.sota2.com/research/sota/road-segmentation-on-worldroadseg-360k.md>): Road Segmentation; WorldRoadSeg 360K; Dice: 78.8.

- [Interactive Road Extraction on WorldRoadSeg-360k](<https://www.sota2.com/research/sota/interactive-road-extraction-on-worldroadseg-360k.md>): Interactive Road Extraction; WorldRoadSeg 360K; Inference Time (ms): 97.31.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2303.11320>)

[PDF](<https://arxiv.org/pdf/2303.11320.pdf>)