TiCLS : Tightly Coupled Language Text Spotter

TL;DR

TiCLS is an end-to-end scene text spotting method that integrates external character-level language models to improve recognition of ambiguous or fragmented text in images.

Contribution

It introduces a linguistic decoder that fuses visual and linguistic features, leveraging pretrained language models for enhanced scene text recognition.

Findings

Achieves state-of-the-art results on ICDAR 2015 and Total-Text datasets.

Demonstrates the effectiveness of external linguistic knowledge in scene text spotting.

Improves recognition robustness for ambiguous or fragmented text instances.

Abstract

Scene text spotting aims to detect and recognize text in real-world images, where instances are often short, fragmented, or visually ambiguous. Existing methods primarily rely on visual cues and implicitly capture local character dependencies, but they overlook the benefits of external linguistic knowledge. Prior attempts to integrate language models either adapt language modeling objectives without external knowledge or apply pretrained models that are misaligned with the word-level granularity of scene text. We propose TiCLS, an end-to-end text spotter that explicitly incorporates external linguistic knowledge from a character-level pretrained language model. TiCLS introduces a linguistic decoder that fuses visual and linguistic features, yet can be initialized by a pretrained language model, enabling robust recognition of ambiguous or fragmented text. Experiments on ICDAR 2015 and Total-Text demonstrate that TiCLS achieves state-of-the-art performance, validating the effectiveness of PLM-guided linguistic integration for scene text spotting.