Tackling the coplanarity problem in 3D camera calibration by means of fuzzy landmarks: a performance study in forensic craniofacial superimposition

Photographic supra-projection is a forensic process that aims to identify a missing person from a photograph and a skull found. Craniofacial superimposition is the second stage of this complex forensic technique devoted to find the most appropriate pose of the skull 3D model to be projected onto the photograph. Craniofacial superimposition can be modeled as a camera calibration problem in computer vision. The process is guided by a number of landmarks identified both in the skull (craniometric landmarks) and in the face (cephalometric landmarks). In this contribution we extend a previous work by studying the influence of the landmark selection procedure in the final superimposition result. In particular, we show how the coplanar landmarks do not carry discriminating depth information in our problem. Moreover, this fact negatively affects the superimposition results. In order to tackle that undesirable situation we propose the use of fuzzy landmarks that could extend the number of original landmarks identified by the forensic expert. Experiments using synthetic and real cases are considered.