A fundamental study on a natural-language-based 3D CG modeling

A method to create a 3D shape based on natural language is proposed. Generally, it is difficult for an non-professional user to create computer graphics (CG) by using traditional CG modelers, so we have developed "Do", a 3D shape modeler utilizing the English language, by which such users would be able to create a CG image. If it were possible to create 3D CG using English by a verbal interface, non-expert users, who do not have any special knowledge of CG, and handicapped users, who are unable to use hardware devices such as a mouse or keyboard, would be able to create CG. However, this system is oriented to shape modeling, and so some scene-constructing functions are attached.