Investigation of Morphological Facial Differences by 3D Averaging

Objectives: To study the feasibility of creating archetypal 3D faces through computerized 3D facial averaging. 3D range finder “Fiore” and its software were used to acquire the 3D scans of the faces while 3D Rugle3 and locally-developed software generated the holistic facial averages. 3D facial averages were created from 2 ethnic groups; European and Japanese and from 3 genetic disorders; Williams syndrome, achondroplasia and Sotos syndrome. Methods: Included averaging corresponding depth (z) coordinates across the 3D facial scans. Compared with other face averaging techniques there was not any warping or “filling in the spaces” by interpolation, however, this facial average lacked color information. Results: Showed that as low as 14 faces were sufficient to create an archetypal facial average. In turn this would make it practical to use face averaging as an identification tool in cases where it would be difficult to recruit a larger number of participants. In generating the average, correcting for size differences among faces was shown to adjust the average outlines of the facial features. Conclusions: 3D facial averaging is expected to help in the identification of ethnic status and in diagnosis of syndromes with distinctive facial features.