Michiko Tsuji1. 1. Section of Maxillofacial Orthognathics, Department of Maxillofacial Reconstruction and Function, Division of Maxillofacial/Neck Reconstruction, Graduate School, Tokyo Medical and Dental University.
Abstract
OBJECTIVES: The purpose of this study was 1) to obtain polygon data and individual parameters from facial photographs of Crouzon syndrome patients and general patients with malocclusion in order to synthesize facial images using a 3 D wire frame model, and 2) to distinguish between Crouzon syndrome patients and general patients with malocclusion using quantitative analysis of characteristic variations of facial features. MATERIALS AND METHODS: The face images were prepared from standardized full-face photographs of 10 Crouzon patients (Crouzon group) and 10 general patients with malocclusion (General group). These facial images were fitted to the 3 D wire frame model and the polygon data were obtained. Following application of principal component analysis, the individual parameters of the standard face were considered to be individual feature parameters. Linear discriminant analysis using the individual feature parameters was applied to distinguish both groups, and the validity of this statistical method was estimated. Moreover, the discriminant rates of the parts of the face were calculated and compared to those of the full face. RESULTS: A discriminant rate of 90% was obtained. The application of individual feature parameters to facial features could be used to objectively distinguish between these two groups. The discriminant rates of the part of the face were lower than those of the full face. CONCLUSION: This method seemed to be effective for diagnosing congenital anomalies from facial features. Further research on other congenital anomalies would be necessary to apply this analysis as a new diagnostic step.
OBJECTIVES: The purpose of this study was 1) to obtain polygon data and individual parameters from facial photographs of Crouzon syndromepatients and general patients with malocclusion in order to synthesize facial images using a 3 D wire frame model, and 2) to distinguish between Crouzon syndromepatients and general patients with malocclusion using quantitative analysis of characteristic variations of facial features. MATERIALS AND METHODS: The face images were prepared from standardized full-face photographs of 10 Crouzon patients (Crouzon group) and 10 general patients with malocclusion (General group). These facial images were fitted to the 3 D wire frame model and the polygon data were obtained. Following application of principal component analysis, the individual parameters of the standard face were considered to be individual feature parameters. Linear discriminant analysis using the individual feature parameters was applied to distinguish both groups, and the validity of this statistical method was estimated. Moreover, the discriminant rates of the parts of the face were calculated and compared to those of the full face. RESULTS: A discriminant rate of 90% was obtained. The application of individual feature parameters to facial features could be used to objectively distinguish between these two groups. The discriminant rates of the part of the face were lower than those of the full face. CONCLUSION: This method seemed to be effective for diagnosing congenital anomalies from facial features. Further research on other congenital anomalies would be necessary to apply this analysis as a new diagnostic step.