Researchers from CREF and the Historical Archive of the Pontifical Gregorian University used a novel approach that combines machine learning and spectroscopic techniques to understand the genesis of an ancient book via handwriting attribution.

The study, published in “Expert System with Applications“, analyzed Clavis Prophetarum, a manuscript by Antonio Vieira, one of the founding fathers of the Portuguese language. The manuscript was previously considered lost and was recently discovered in the Historical Archives of the Pontifical Gregorian University in Rome.

The complexity of the original book presented an exciting challenge as it was written over a period of 50 years through corrections, revisions, censorship and annotations by various authors.

Researchers from “Laboratorio di fisica per i beni culturali” Giulia Festa, Lorenzo Teodonio, and Claudia Scatigno,  used a combination of Machine Learning and X-ray fluorescence and Fourier transform infrared spectroscopies to recognize the different inks used in the manuscript and identify the authors.

This innovative methodology has great potential for the authentication of both ancient and modern manuscripts and for handwriting analysis.

Read the full paper here