Open-set Text Recognition via Character-Context Decoupling

TL;DR

This paper introduces a character-context decoupling framework for open-set text recognition, effectively separating visual and contextual information to improve recognition of novel characters in various scenarios.

Contribution

The paper proposes a novel framework that decouples contextual and visual information, including temporal and linguistic aspects, to enhance open-set text recognition performance.

Findings

Achieves promising results on open-set, zero-shot, and close-set datasets.

Effectively separates temporal and linguistic information from visual features.

Improves recognition accuracy for novel characters in open-set scenarios.

Abstract

The open-set text recognition task is an emerging challenge that requires an extra capability to cognize novel characters during evaluation. We argue that a major cause of the limited performance for current methods is the confounding effect of contextual information over the visual information of individual characters. Under open-set scenarios, the intractable bias in contextual information can be passed down to visual information, consequently impairing the classification performance. In this paper, a Character-Context Decoupling framework is proposed to alleviate this problem by separating contextual information and character-visual information. Contextual information can be decomposed into temporal information and linguistic information. Here, temporal information that models character order and word length is isolated with a detached temporal attention module. Linguistic information that models n-gram and other linguistic statistics is separated with a decoupled context anchor mechanism. A variety of quantitative and qualitative experiments show that our method achieves promising performance on open-set, zero-shot, and close-set text recognition datasets.