Reconstructing facial features from skeletal remains using CT scans

Identifying skeletal remains is a challenging problem that requires both technical knowledge of bone and face structure as well as creative anatomical artistry. Traditionally this task has been undertaken by trained specialists who spend hours carving or drawing potential faces for criminal investigations, historical explorations, and disaster relief recordings.

Deep Neuron CT Facial Reconstruction aims to apply generative neural networks to automatically create realistic looking faces from skulls with no human interaction and in real-time. By using a dataset of 3D head CT scans, the team is exploring both 2D and 3D generative deep learning techniques to create a proof of concept.