Facial recognition as a tool to identify Roman emperors: towards a new methodology

Portraits of Roman emperors are traditionally recognised by their unique coiffure patterns, a method that runs the risk of ignoring portraits that do not cohere to the standardised image of the emperor. This article investigates whether it is possible to recognise and distinguish emperors using the facial features of their portraits. By using a technique called transfer learning, it utilises existing deep-learning facial recognition models, augmented with images of Roman imperial portraits, to provide a new empirical foothold in the debate of Roman emperor recognition. The results of the experiments demonstrate that by only a limited amount of training, such a so-called "pre-trained" model (i.e., InceptionResnet-V1) is able to correctly classify most images in the dataset of Roman emperors. As such, this article has made a first step towards applying facial recognition models to the study of ancient imperial portraiture.

Darshan Srirangachar Ramesh, Sam Heijnen, Olivier Hekster, Luuk Spreeuwers, Florens de Wit