REALIZATION OF COMPUTER-AIDED ENVIRONMENTS FOR ORIGINAL THINKING BASING UPON RECOGNITION AND GENERATION OF THREE DIMENSIONAL IMAGES, AND EXPERIENCING OF VIRUTUAL WORLD

在这项研究中，我们开发了一些工具，用于支持的概念与高水平的原创性，使用三维（3D）图像作为媒体的推导。这里，3D图像是指存储在3D体素阵列中的数字化3D图片。（1）识别研究了体素型三维灰度图像的内部结构提取算法。包括灰度图像的细化、基于在3D图像上计算的集中指数的特征提取、弯曲的3D表面上的突起的提取、3D差分滤波器等。（2）三维图像处理的视觉专家系统。我们开发的视觉专家系统称为IMPRESS的2D图像扩展到3D图片。这道题的难点是输入样图的方法。我们开发了一些替代方案，包括一个草图上的MIP图像和一个快速和近似版本的体积渲染。（3）用手在真实的世界中操纵3D虚拟空间中的虚拟对象的模型。设计了一种带传感器的手套式虚拟物体抓取、搬运和堆放装置，并进行了实际测试。

In this study, we developed a few tools for supporting the derivation of conception with high level of originality using three dimensional (3D) images as the media. Here a 3D image means a digitized 3D picture stored in a 3D voxel array.(1)Recognition Several algorithms for extraction of inner structure from voxel type 3D gray picutre were developed. Thinning of gray pictures, feature extarction based on the concentration index calculated on a 3D picture, extraction of a protrusion on a curved 3D surface, 3D difference filter, etc. are included.(2)Vision expert system for 3D picture processing. The vision expert system called IMPRESS which we developed for a 2D image was extended to a 3D picture. Difficulty in this problem is the method to input a sample figure. We developed a few alternatives including a sketch on a MIP image and a fast and approximated version of the volume rendering.(3)Model for manipulating virtual objects in 3D virtual space with the hand in real world. Using a glove-like device with sensors to measure angles among parts of fingers, grasping, carrying and stacking virtual objects were developed, and tested practically.