Impact of geometry and viewing angle on classification accuracy of 2D based analysis of dysmorphic faces

• Published in: European journal of medical genetics Vol. 51; no. 1; pp. 44 - 53
• Main Authors: Vollmar, Tobias, Maus, Baerbel, Wurtz, Rolf P, Gillessen-Kaesbach, Gabriele, Horsthemke, Bernhard, Wieczorek, Dagmar, Boehringer, Stefan
• Format: Journal Article Web Resource
• Language: English
• Published: Amsterdam Elsevier Masson SAS 01.01.2008; Elsevier Science; Elsevier
• Subjects: Abnormalities, Multiple - pathology; Adolescent; Adult; Artificial Intelligence; Automated pattern recognition; Biological and medical sciences; Child; Child, Preschool; Classification; Computer-assisted diagnosis; Congenital Abnormalities - pathology; Craniofacial Abnormalities - pathology; Dysmorphism; Face; Face - abnormalities; Female; Fundamental and applied biological sciences. Psychology; General aspects. Genetic counseling; genetic processes; Genetics; Genetics & genetic processes; Genetics of eukaryotes. Biological and molecular evolution; Génétique & processus génétiques; Humans; Image Interpretation, Computer-Assisted; Infant; Life sciences; Male; Medical Education; Medical genetics; Medical sciences; Middle Aged; Sciences du vivant; Software; Syndrome; Automated pattern recognition; Dysmorphism; Computer-assisted diagnosis; Face; Classification; Human; Automation; Method; Accuracy; Malformation; Genetics; Diagnostic aid; Recognition
• ISSN: 1769-7212; 1878-0849; 1878-0849
• DOI: 10.1016/j.ejmg.2007.10.002

Abstract Digital image analysis of faces has been demonstrated to be effective in a small number of syndromes. In this paper we investigate several aspects that help bringing these methods closer to clinical application. First, we investigate the impact of increasing the number of syndromes from 10 to 14 as compared to an earlier study. Second, we include a side-view pose into the analysis and third, we scrutinize the effect of geometry information. Picture analysis uses a Gabor wavelet transform, standardization of landmark coordinates and subsequent statistical analysis. We can demonstrate that classification accuracy drops from 76% for 10 syndromes to 70% for 14 syndromes for frontal images. Including side-views achieves an accuracy of 76% again. Geometry performs excellently with 85% for combined poses. Combination of wavelets and geometry for both poses increases accuracy to 93%. In conclusion, a larger number of syndromes can be handled effectively by means of image analysis.