ConText: Driving In-context Learning for Text Removal and Segmentation

TL;DR

This paper introduces ConText, a novel visual in-context learning approach for text removal and segmentation that enhances reasoning through task chaining, context-aware aggregation, and self-prompting, achieving state-of-the-art results.

Contribution

The paper proposes a new task-chaining compositor and context-aware aggregation techniques to improve in-context learning for optical character recognition tasks.

Findings

Achieves state-of-the-art performance on multiple benchmarks.

Effectively handles visual heterogeneity with self-prompting.

Enhances reasoning with task chaining and context integration.

Abstract

This paper presents the first study on adapting the visual in-context learning (V-ICL) paradigm to optical character recognition tasks, specifically focusing on text removal and segmentation. Most existing V-ICL generalists employ a reasoning-as-reconstruction approach: they turn to using a straightforward image-label compositor as the prompt and query input, and then masking the query label to generate the desired output. This direct prompt confines the model to a challenging single-step reasoning process. To address this, we propose a task-chaining compositor in the form of image-removal-segmentation, providing an enhanced prompt that elicits reasoning with enriched intermediates. Additionally, we introduce context-aware aggregation, integrating the chained prompt pattern into the latent query representation, thereby strengthening the model's in-context reasoning. We also consider the issue of visual heterogeneity, which complicates the selection of homogeneous demonstrations in text recognition. Accordingly, this is effectively addressed through a simple self-prompting strategy, preventing the model's in-context learnability from devolving into specialist-like, context-free inference. Collectively, these insights culminate in our ConText model, which achieves new state-of-the-art across both in- and out-of-domain benchmarks. The code is available at https://github.com/Ferenas/ConText.