Tracking the digital dead:  A progress report in the use of transfer learning, image analysis, and forensic anthropology to classify human remains sold on social media

1. Abstract

This short presentation presents preliminary results of new work expanding on that of the ongoing Bone Trade Project (https://bonetrade.github.io/). It discusses progress in the use of transfer learning to create a hand-held application to classify human remains observed for sale online. Global legal loopholes allow the trade, sale, or auction of human remains of questionable provenance and provenience, creating the need to better map trafficking across platforms and understand its socio-cultural context as a community of practice. Uniting osteology, forensic anthropology and image analysis of large corpuses from public Facebook, Instagram, etc., this presentation demonstrates the potential to monitor this trade in real time. A progress report of mobile app development will be presented, as well as ongoing and future technical and ethical issues that have surfaced to date. Ultimately, we demonstrate how the technologies of surveillance can be subverted to restore some measure of humanity to the commodified dead.