Advancing Post-OCR Correction: A Comparative Study of Synthetic Data

TL;DR

This study evaluates synthetic data's role in post-OCR correction, introducing a new glyph similarity algorithm and demonstrating improved model performance, especially in low-resource languages, without manual annotations.

Contribution

We propose a novel glyph similarity algorithm for synthetic data generation and demonstrate its effectiveness in improving OCR correction across multiple languages.

Findings

Synthetic data reduces Character Error Rates in OCR models.

Our method outperforms traditional synthetic data generation techniques.

Significant improvements observed in low-resource language OCR correction.

Abstract

This paper explores the application of synthetic data in the post-OCR domain on multiple fronts by conducting experiments to assess the impact of data volume, augmentation, and synthetic data generation methods on model performance. Furthermore, we introduce a novel algorithm that leverages computer vision feature detection algorithms to calculate glyph similarity for constructing post-OCR synthetic data. Through experiments conducted across a variety of languages, including several low-resource ones, we demonstrate that models like ByT5 can significantly reduce Character Error Rates (CER) without the need for manually annotated data, and our proposed synthetic data generation method shows advantages over traditional methods, particularly in low-resource languages.