Text Detection Forgot About Document OCR

TL;DR

This paper compares in-the-wild and document-specific text recognition methods, showing that in-the-wild methods perform competitively on structured documents and emphasizing the importance of including document OCR in evaluations.

Contribution

It provides a comprehensive comparison of in-the-wild and document OCR methods on structured documents, highlighting the relevance of in-the-wild methods for document text detection.

Findings

In-the-wild text detection methods perform well on document images.

State-of-the-art in-the-wild methods outperform traditional OCR in some cases.

Evaluation of text detection should include both in-the-wild and document-specific methods.

Abstract

Detection and recognition of text from scans and other images, commonly denoted as Optical Character Recognition (OCR), is a widely used form of automated document processing with a number of methods available. Yet OCR systems still do not achieve 100% accuracy, requiring human corrections in applications where correct readout is essential. Advances in machine learning enabled even more challenging scenarios of text detection and recognition "in-the-wild" - such as detecting text on objects from photographs of complex scenes. While the state-of-the-art methods for in-the-wild text recognition are typically evaluated on complex scenes, their performance in the domain of documents is typically not published, and a comprehensive comparison with methods for document OCR is missing. This paper compares several methods designed for in-the-wild text recognition and for document text recognition, and provides their evaluation on the domain of structured documents. The results suggest that state-of-the-art methods originally proposed for in-the-wild text detection also achieve competitive results on document text detection, outperforming available OCR methods. We argue that the application of document OCR should not be omitted in evaluation of text detection and recognition methods.