A method for quantifying facial muscle movements in the smile during facial expression training
Ai Takami, Kyoko Ito, Shogo Nishida · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · 2008
The purpose of this study is to propose an evaluation method capable of quantifying facial expressions during facial expression training that is intended to achieve a more expressive face. The specific aim was to investigate methods of estimating facial muscle movements from facial images and display our estimation results in an understandable way, as well as to evaluate the effectiveness of these methods. The facial expression that our study chose to work with was the smile. Facial muscles around the mouth that need to contract when a person creates a smile were selected. The muscle parameters that indicate a muscle's level of contraction were calculated using the distances of various feature points in facial images. Further, a schematic showing quasi-muscles was overlaid upon a facial image to show the estimation results. An evaluation experiment was conducted. The accuracy of muscle parameters derived from facial expressions that change greatly from a face devoid of expression was higher than that of facial expressions that changed very little. From the questionnaire results, it was determined that the quasi-muscles on a facial image were found to be helpful in understanding the meaning of muscle parameters.