Abstract
The Kansei engineering is a method to translate Kansei and images of human being into physical factors of design in order to design products, environment, etc. which meet the Kansei. Utilizing this technology is considered to be significant in the fields of food or environment in future with the following reasons: if Kansei can be comprehended, it will be possible for agricultural or food products that are highly evaluated by consumers to be designed, developed, and evaluated. Likewise, for comfortable room space, city and rural landscape with high amenity effects to be constructed. So far we have focused on the relation between facial expression and Kansei, and we have specified significant face nodes and constructed the 2-dimensional facial expression - Kansei estimation system using degree of similarity and the nearest neighbor rule. This system has implied the possibility that Kansei is presumable from expression of the face. In this paper, extending from 2-dimensional, we constructed the 3-dimensional facial expression - Kansei estimation system using stereovision and made an evaluation. When we evaluated for the object of positive Kansei, negative Kansei, and surprise, we obtained good results particularly for the positive Kansei and negative Kansei. As for surprise, the discernment result was lower than the other two Kansei because of individual differences, expressional differences of Kansei, or etc. But we consider that this system indicates the possibility of Kansei to be estimated.
