The OCR-PT-CT Project: Semi-Automatic Recognition of Ancient Egyptian Hieroglyphs Based on Metric Learning

TL;DR

This paper presents a semi-automatic recognition system for ancient Egyptian hieroglyphs that combines a neural network and a novel deep metric learning approach, improving accuracy and adaptability for digital humanities research.

Contribution

It introduces a deep metric learning method for hieroglyph recognition that outperforms traditional neural networks, especially with limited or imbalanced data.

Findings

Deep Metric Learning achieves 97.70% accuracy.

Neural network with MobileNet reaches 93.87% accuracy.

System effectively recognizes hieroglyphs from Coffin and Pyramid Texts.

Abstract

Digital humanities are significantly transforming how Egyptologists study ancient Egyptian texts. The OCR-PT-CT project proposes a recognition method for hieroglyphs based on images of Coffin Texts (CT) from Adriaan de Buck (1935-1961) and Pyramid Texts (PT) from Middle Kingdom coffins (James Allen, 2006). The system identifies hieroglyphs and transcribes them into Gardiner's codes. A web tool organizes them by spells and witnesses, storing the data in CSV format for integration with the MORTEXVAR dataset, which collects Coffin Texts with metadata, transliterations, and translations for research. Recognition has been addressed in two ways: a Mobilenet neural network trained on 140 hieroglyph classes achieved 93.87 \% accuracy but struggled with underrepresented classes. A novel Deep Metric Learning approach improves flexibility for new or data-limited signs, achieving 97.70 \% accuracy and recognizing more hieroglyphs. Due to its superior performance under class imbalance and adaptability, the final system adopts Deep Metric Learning as the default classifier.