What question did this study set out to answer?

This research aims to analyze micro-CT data to enhance the detection of ink patterns in Herculaneum papyri.

March 18, 2026Open Access

Novel Findings in Ink Detection for Carbonized Herculaneum Papyri Using Machine Learning

Key Points

This research aims to analyze micro-CT data to enhance the detection of ink patterns in Herculaneum papyri.
Applied machine learning methods to analyze micro-CT data.
Utilized datasets from the Vesuvius Challenge for training models.
Investigated structural patterns in ink signals under various scanning parameters.
Identified previously unreported structural patterns in ink signals.
Showed variation in ink patterns based on different scanning parameters.
Introduced a new training methodology that improved model performance.

Abstract

This technical report presents novel findings from computational analysis of micro-CT data of carbonized Herculaneum papyrus fragments. Using machine learning methods applied to publicly available datasets from the Vesuvius Challenge, we identify previously unreported structural patterns in ink signals, demonstrate that these patterns vary with scanning parameters, and introduce a training methodology that improves model performance. The package contains a full manuscript, supporting figures, and numerical data. Detailed methodology will be released upon formal publication.Data source: EduceLab-Scrolls Dataset (Parsons, S., Parker, C.S., Chapman, C., Hayashida, M., & Seales, W.B., 2023. EduceLab-Scrolls: Verifiable Recovery of Text from Herculaneum Papyri using X-ray CT. arXiv:2304.02084). Licensed under CC BY-NC 4.0.

Novel Findings in Ink Detection for Carbonized Herculaneum Papyri Using Machine Learning

Key Points

Abstract

Cite This Study