Abstract
While traders of human remains on Instagram will give some indication, their best estimate, or repeat hearsay, regarding the geographic origin or provenance of the remains, how can we assess the veracity of these claims when we cannot physically examine the remains? A novel image analysis using convolutional neural networks in a one-shot learning architecture with a triplet loss function is used to develop a range of ‘distances’ to known ‘reference’ images for a group of skulls with known provenances and a group of images of skulls from social media posts. Comparing the two groups enables us to predict a broad geographic ‘ancestry’ for any given skull depicted, using a mixture discriminant analysis, as well as a machine-learning model, on the image dissimilarity scores. It thus seems possible to assign, in broad strokes, that a particular skull has a particular geographic ancestry.
Highlights
The forensic determination of ‘ancestry’ and estimation of sex of unprovenienced human remains relies on the careful measurement of ‘landmarks’ on the elements at hand that have been found to be diagnostic when suitably calibrated, standardized for population, and analyzed
In this paper we explore the potential for machine vision on simple photographs of human skulls as an initial experiment that uses a particular kind of machine vision architecture to develop a suite of ‘distances’ from known reference images, and perform a mixture discriminant analysis comparing a dataset with known, grounded provenience against a dataset sourced from social media posts
In other research we showed that there were ethical and technological problems with using neural networks to classify these images of human remains, or in using transfer-learning techniques which require thousands of images of a particular classification in order to work (Huffer & Graham 2018; Huffer, Wood & Graham 2019)
Summary
The forensic determination of ‘ancestry’ and estimation of sex of unprovenienced human remains (i.e. remains for which the archaeological origin is unknown) relies on the careful measurement of ‘landmarks’ on the elements at hand (ideally crania and/or pelves) that have been found to be diagnostic when suitably calibrated, standardized for population, and analyzed. Researchers have begun to apply machine learning techniques to this data, with very good results (e.g. Navega et al 2015a, 2015b; Ousley 2016; Maier et al 2015) suggesting the potential to improve the accuracy of identifying unprovenienced remains at the population and individual levels This is a critical question to investigate because traders buy and sell human remains online; there is a very active market for human bone, and to date it is nearly impossible to say anything about which people(s) are being bought and sold (see section 1.2). How many people’s remains are being bought and sold? From what areas of the world (or populations) have their remains been sourced? We believe we can begin to answer such previously imponderable questions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
