Post-OCR Text Correction for Bulgarian Historical Documents

TL;DR

This work creates the first benchmark dataset for evaluating the OCR text correction for historical Bulgarian documents written in the first standardized Bulgarian orthography: the Drinov orthography from the 19th century, and develops a method for automatically generating synthetic data in this orthography by leveraging vast amounts of contemporary literature Bulgarian texts.

Abstract

The digitization of historical documents is crucial for preserving the cultural heritage of the society. An important step in this process is converting scanned images to text using Optical Character Recognition (OCR), which can enable further search, information extraction, etc. Unfortunately, this is a hard problem as standard OCR tools are not tailored to deal with historical orthography as well as with challenging layouts. Thus, it is standard to apply an additional text correction step on the OCR output when dealing with such documents. In this work, we focus on Bulgarian, and we create the first benchmark dataset for evaluating the OCR text correction for historical Bulgarian documents written in the first standardized Bulgarian orthography: the Drinov orthography from the 19th century. We further develop a method for automatically generating synthetic data in this orthography, as well as in the subsequent Ivanchev orthography, by leveraging vast amounts of contemporary literature Bulgarian texts. We then use state-of-the-art LLMs and encoder-decoder framework which we augment with diagonal attention loss and copy and coverage mechanisms to improve the post-OCR text correction. The proposed method reduces the errors introduced during recognition and improves the quality of the documents by 25\%, which is an increase of 16\% compared to the state-of-the-art on the ICDAR 2019 Bulgarian dataset. We release our data and code at \url{https://github.com/angelbeshirov/post-ocr-text-correction}.}

Figures (5)

• Figure 1: Pipeline for post-OCR text correction.
• Figure 2: Example sentence from the ICDAR 2019 dataset.
• Figure 3: Example of a synthetically generated sentence pair in the Ivanchev orthography. The first sentence is correct, while the second is misspelled.
• Figure 4: Distribution of the normalized Levenshtein distance for the ICDAR 2019 and DOPOC datasets.
• Figure 5: Our model manages to fix most of the errors, as shown in the examples from the first two columns. However, in some cases it fails to properly correct the error, as demonstrated in the two subsequent columns.