STEP -- Towards Structured Scene-Text Spotting

TL;DR

This paper introduces the structured scene-text spotting task and proposes STEP, a model that dynamically detects and recognizes text based on user-defined regular expressions, enabling accurate zero-shot OCR in complex real-world scenarios.

Contribution

The paper presents a novel structured scene-text spotting task and a model, STEP, that conditions OCR on regular expressions, handling spaces and various granularities, trained solely on public data.

Findings

STEP achieves accurate zero-shot text spotting in diverse scenarios.

The new dataset includes out-of-vocabulary structured text like prices and serial numbers.

STEP outperforms baseline methods in structured text recognition tasks.

Abstract

We introduce the structured scene-text spotting task, which requires a scene-text OCR system to spot text in the wild according to a query regular expression. Contrary to generic scene text OCR, structured scene-text spotting seeks to dynamically condition both scene text detection and recognition on user-provided regular expressions. To tackle this task, we propose the Structured TExt sPotter (STEP), a model that exploits the provided text structure to guide the OCR process. STEP is able to deal with regular expressions that contain spaces and it is not bound to detection at the word-level granularity. Our approach enables accurate zero-shot structured text spotting in a wide variety of real-world reading scenarios and is solely trained on publicly available data. To demonstrate the effectiveness of our approach, we introduce a new challenging test dataset that contains several types of out-of-vocabulary structured text, reflecting important reading applications of fields such as prices, dates, serial numbers, license plates etc. We demonstrate that STEP can provide specialised OCR performance on demand in all tested scenarios.