Multi-modal In-Context Learning Makes an Ego-evolving Scene Text Recognizer

TL;DR

This paper introduces E$^2$STR, a scene text recognition model trained with context-rich sequences to enable effective in-context learning, outperforming fine-tuned methods without additional training.

Contribution

It proposes a novel in-context training strategy for STR, allowing regular-sized models to perform well in diverse scenarios without fine-tuning.

Findings

E$^2$STR achieves superior performance on public benchmarks.

The model demonstrates remarkable training-free adaptation.

It outperforms state-of-the-art fine-tuned approaches.

Abstract

Scene text recognition (STR) in the wild frequently encounters challenges when coping with domain variations, font diversity, shape deformations, etc. A straightforward solution is performing model fine-tuning tailored to a specific scenario, but it is computationally intensive and requires multiple model copies for various scenarios. Recent studies indicate that large language models (LLMs) can learn from a few demonstration examples in a training-free manner, termed "In-Context Learning" (ICL). Nevertheless, applying LLMs as a text recognizer is unacceptably resource-consuming. Moreover, our pilot experiments on LLMs show that ICL fails in STR, mainly attributed to the insufficient incorporation of contextual information from diverse samples in the training stage. To this end, we introduce E$^2$STR, a STR model trained with context-rich scene text sequences, where the sequences are generated via our proposed in-context training strategy. E$^2$STR demonstrates that a regular-sized model is sufficient to achieve effective ICL capabilities in STR. Extensive experiments show that E$^2$STR exhibits remarkable training-free adaptation in various scenarios and outperforms even the fine-tuned state-of-the-art approaches on public benchmarks. The code is released at https://github.com/bytedance/E2STR .