Karl-Friedrich Krey1, Karl-Heinz Dannhauer. 1. Poliklinik für Kieferorthopädie, Zentrum für Zahn-, Mund- und Kieferheilkunde, Universität Leipzig, Leipzig, Germany. krek@medizin.uni-leipzig.de
Abstract
AIM: Are centroid size, principal component analysis (PCA) and thin-plate splines (TPS) sufficient for determining facial morphology? Is it possible to identify relationships between cranial morphology, gender and facial profile? MATERIAL AND METHODS: Profile photos of 110 adult patients were measured according to A.M. Schwarz landmarks using Onyx Ceph. Centroid size was calculated from x and y-coordinates. After Procrustes transformation, a principal component analysis for identifying major components of facial morphology was performed and the results visualized using thin-plate splines. At the same time, lateral cephalograms of all patients were analyzed according to Hasund. RESULTS: There were significant differences in centroid size betweeen male and female patients. Only the vertical skeletal structure had an impact on centroid size. Six components (PC1 to PC6) were identified using PCA. They were responsible for 86.5% of the variance. PC1 (33.9%) described scaling along an axis from Porion to chin. PC2 (28.6%) characterized the vertical dimensions of the lower face. Significant differences were only apparent between males and females in PC3 and PC4. In terms of cephalometric parameters, PC2 and PC3 differed in the vertical, and PC1 und PC2 in the sagittal configuration. CONCLUSIONS: The analyses presented here suffice for describing facial morphology qualitatively and quantitatively as demonstrated by this example. Separating size from shape is useful for investigating therapeutically and growth-related morphological changes. It is difficult to draw conclusions about skeletal parameters.
AIM: Are centroid size, principal component analysis (PCA) and thin-plate splines (TPS) sufficient for determining facial morphology? Is it possible to identify relationships between cranial morphology, gender and facial profile? MATERIAL AND METHODS: Profile photos of 110 adult patients were measured according to A.M. Schwarz landmarks using Onyx Ceph. Centroid size was calculated from x and y-coordinates. After Procrustes transformation, a principal component analysis for identifying major components of facial morphology was performed and the results visualized using thin-plate splines. At the same time, lateral cephalograms of all patients were analyzed according to Hasund. RESULTS: There were significant differences in centroid size betweeen male and female patients. Only the vertical skeletal structure had an impact on centroid size. Six components (PC1 to PC6) were identified using PCA. They were responsible for 86.5% of the variance. PC1 (33.9%) described scaling along an axis from Porion to chin. PC2 (28.6%) characterized the vertical dimensions of the lower face. Significant differences were only apparent between males and females in PC3 and PC4. In terms of cephalometric parameters, PC2 and PC3 differed in the vertical, and PC1 und PC2 in the sagittal configuration. CONCLUSIONS: The analyses presented here suffice for describing facial morphology qualitatively and quantitatively as demonstrated by this example. Separating size from shape is useful for investigating therapeutically and growth-related morphological changes. It is difficult to draw conclusions about skeletal parameters.
Authors: Jelena Djordjevic; Debbie A Lawlor; Alexei I Zhurov; Arshed M Toma; Rebecca Playle; Stephen Richmond Journal: BMJ Open Date: 2013-05-28 Impact factor: 2.692