3D semi-landmarks-based statistical face reconstruction

The aim of craniofacial reconstruction is to estimate the shape of a face from the shape of the skull. Few works in machine-assisted facial reconstruction have been conducted so far, probably due to technical poor machine performance and data availability and theoret-ical complexity reasons. Therefore, the main works in the literature consist in manual reconstructions. In this paper, an original approach is first proposed to build a 3D statistical model of the skullface set from 3D CT scans. Then, a reconstruction method is introduced in order to estimate, from this statistical model, the 3D facial shape of one subject from known skull data