In silico modeling of smiling motions -aging in adulthood

in silico modeling of smiling motions (maximum lip corner retraction) was achieved using the recorded 3-dimensional motion data of 10 anatomical landmarks on a face for 60 women (30 young adults and 30 middle-aged adults). A total of 55 feature variables were extracted from the motion data to generate a feature vector. Sets of the feature vector and the age category were stored as knowledge in the model. The prediction of the age category (young adult or middle-aged adult) was performed by means of Nm-neighboring search in the model knowledge with the weight coefficient W. The prediction model was successfully optimized to provide the prediction accuracy of 75.3% and the feature elements to well represent the effect of aging on smiling motions were formulated objectively.