Optical Character Recognition of 19th Century Classical Commentaries: the Current State of Affairs

TL;DR

This paper evaluates OCR performance on 19th-century classical commentaries, demonstrating that Kraken + Ciaconna outperforms Tesseract/OCR-D on Greek texts, and releases datasets and models for improved OCR of historical manuscripts.

Contribution

The study compares OCR pipelines for classical commentaries and provides new datasets and pre-trained models for ancient Greek OCR tasks.

Findings

Kraken + Ciaconna achieves lower CER on Greek sections.

Tesseract/OCR-D performs slightly better on Latin script sections.

New datasets and models are released for OCR of ancient Greek texts.

Abstract

Together with critical editions and translations, commentaries are one of the main genres of publication in literary and textual scholarship, and have a century-long tradition. Yet, the exploitation of thousands of digitized historical commentaries was hitherto hindered by the poor quality of Optical Character Recognition (OCR), especially on commentaries to Greek texts. In this paper, we evaluate the performances of two pipelines suitable for the OCR of historical classical commentaries. Our results show that Kraken + Ciaconna reaches a substantially lower character error rate (CER) than Tesseract/OCR-D on commentary sections with high density of polytonic Greek text (average CER 7% vs. 13%), while Tesseract/OCR-D is slightly more accurate than Kraken + Ciaconna on text sections written predominantly in Latin script (average CER 8.2% vs. 8.4%). As part of this paper, we also release GT4HistComment, a small dataset with OCR ground truth for 19th classical commentaries and Pogretra, a large collection of training data and pre-trained models for a wide variety of ancient Greek typefaces.