The Development of the Technique Objectively Estimating Impressions of Facial Configuration

To perform make-up appreciating the features of individual face, the make-up technique abstracted the attractiveness of the recognized facial features is important. In this study, we tried to elucidate the important factors when we recognize the pattern of other's entire face, and to develop the method that objectively interprets the factors from a face image. The face photographs of Japanese women in their twenties were used in the experiments. The first experiment was done to find out the important factors in the recognition of faces. The subjects were given the task to classify 112 facial photographs based on the criterion of “resemblance.” The result of cluster analysis showed that faces were classified into 5 categories, and it was elucidated that the impressions such as “the size of the eyes” and “the contour of the face below the eyes” are essential factors in categorizing the face. In the second experiment, the subjects evaluated the impression of the same photographs based on the two factors. The multiple regression analysis was applied to 57 measured values in the face image as explanatory variable and impression values obtained by subjects' evaluation as criterion variable. As a result, the impression value was estimated by the selected 10 measured values in the face image. Furthermore, we developed a facial configuration map on which the face image can be plotted by the axes of two factors. Face images plotted onto the map were classified into the 5 categories at an accuracy rate of nearly 80%, thus confirming the validity that two factors were essential in recognizing faces. The facial configuration map made it possible to objectively assess facial features. We think these results contribute to develop the make-up technique appreciating the features of individual face.