A Computer Based Objective Grading System for Facial Paralysis

A computer assisted objective grading system based on the asymmetry analysis of patient threedimensional surface models obtained from a 3D facial imaging system is developed for facial paralysis diagnosis. To overcome the subjectivity of traditional diagnosis methods, facial asymmetry grading is carried out according to fine registration result of the original and mirror facial meshes which do not rely on any landmarks. As a higher order property of 3D surfaces, the Principal, Gaussian and Mean Curvatures, Shape Index and Curvedness are introduced for grading the asymmetry of the faces as descriptions of the surface local features besides examining the local differences in distance between the original and mirrored surfaces. The present results show that the proposed approach is able to objectively assess the asymmetrical of human faces as well as to detect the improvement of the patients quite well.