Statistical shape model (SSM) black South African faces based on from facial landmarks on a CT and CBCT scan population

The research developed normative reference values of black South African faces for various inter-landmark distances, and derived a statistical shape model (SSM) of 3D facial shape variation which can be applied to estimate missing soft tissue segments on simulated defective faces, such as for facial prosthetics. The research was conducted on 235 computed tomography and cone beam computed tomography scans.

1) SSM_BlackSA: The statistical shape model containing the shape variation of a black South African adult population is provided as a matlab output file

2) Defect estimation accuracy: an excel file containing the root mean square errors for estimating defect reconstruction accuracy for various facial defects, including for the full nose, partial nose, orbit, cheek, lips, bi-ortibal and two types of composite defects (small and large)

3) Intra and interobserver distances: this excel provides data of 2 observers that each manually calculated inter-landmark distances for 20 clinically relevant distances on the face. Each observer had 3 sets of landmarks.

4) Landmark_coordinates: This excel supplies the x,y,z coordinates of 24 facial landmarks of 235 sample faces.

5) Manual_interlandmark_distances: this excel gives the calculated inter-landmark measurements of 20 clinically relevant distances calculated from manually placed landmarks on the face. This information was used to determine the population norms.

6) Samples_list: an excel with the samples used in this research with basic demographic information such as sex, age and modality

7) SSM_Graphic: a figure showing the visual representation of the SSM provided as a matlab file. This figure indicates the modes of variation most responsible for variation, as well as the multinormal distribution of variation