Development of Design Support System Applying Visual Language on VR

A)曲线和曲面的研究在游戏和动画作品中，实体模型（我们称之为“人物”）和三维计算机图形模型通常是在二维人物的基础上制作的。对人脸实体模型制作技术的需求越来越大，但作为该技术基础的建模模式还很模糊。然后，我们以人偶、日本人偶和licca人偶为例，本研究的目的是通过实例制作人物设计辅助系统，并研究三维模型与二维图像之间的面部轮廓关系。首先，我们清除了娃娃面部曲线之间的差异，以及它们是如何从人脸中抽象出来的。其次，我们研究了日本玩偶和浮世绘之间的面部轮廓差异，以及一些动画作品中人物和2D人物之间的面部轮廓差异。第三，我们制作了设计辅助系统，该系统利用所获得的人偶特征，输出人偶面部的七条横截线，这些横截线反映了任意一个人偶的面部特征。B)设计支持系统的研究很少有研究解决从汽车印象中提取汽车设计中最重要的最优配置项的逆问题。有必要应用非线性模型来处理有关这种感性的问题。然而，早期的研究很少使用粗糙集来处理非线性问题。因此，在本研究中，我们将粗糙集作为求解逆问题的方法。本研究的目的是利用粗糙集来阐明非线性最优解的特征，并通过实验设计将其与线性最优解进行比较。具体来说，我们通过实验设计比较了高覆盖指标值的决策规则条件部分与高覆盖指标值的主效应。此外，我们考虑了两种方法的特点，并能够澄清每种方法的特点。最后，我们开发了一个系统，通过可视化两种最优解决方案来支持设计。少

A)Research of curves and surfacesSolid models (we call "figures") and 3D computer graphics models are often made on the basis of 2D characters in games and animation works. Demand of the technique about how to make solid model of human face is increasing, but those modelling patterns as a base of the technique is vague yet.Then, we took figures, Japanese dolls, and licca dolls as the examples, and the aim of this study was to make the character design assistance system by the examples and to research the relation of face's outline between 3D model and 2D picture. First, we cleared the differences among curves of the dolls' faces and how they were abstracted from human faces. Second, we researched the differences of face's outline between Japanese dolls and ukiyoe, and between figures and 2D characters in some animation works. Third, we made the design assistance system, which outputted seven cross-sectional lines of doll's face reflecting a optional person face's features, by the use o … More f the dolls' features obtained.B)Research of design support systemsA few researches solving inverse problems that extracted optimum configuration items that were the most important in car design from car's impressions. It is necessary to apply a nonlinear model for treating the problem concerning such a kansei. However, few earlier researches used rough set that could treat non-liner problem. Therefore, in this study, we applied rough set as methods for solving inverse problems. The aim of this study is to clarify the features of non-linear optimum solutions by using rough set and compared those solutions with linear optimum solutions by design of experiments. Concretely, we compared condition parts of decision rules with a high covering index value with main effect with a high value by design of experiments. In addition we considered the feature of both methods, and were able to clarify the features of each. Finally, we developed the system that supported the design by visualizing both optimum solutions. Less