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

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