1990 Fiscal Year Final Research Report Summary
Research on a Prototype System to Integrate and Understand Plural Information
| Project/Area Number |
63850069
|
|---|
| Research Category |
Grant-in-Aid for Developmental Scientific Research (B).
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
計算機工学
|
|---|
| Research Institution | Waseda Univ. |
|---|
Principal Investigator |
OHARA Hiroyoshi Waseda Univ. S&E Prof., 理工学部, 教授 (40063367)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ITOH Yukihiro Shizuoka Univ. Eng. Associate Prof., 工学部, 助教授 (20193526)
TAKAGI Akira CSK Corp. Department Manager, 部長
SHIRAI Katsuhiko Waseda Univ. S&E Prof., 理工学部, 教授 (10063702)
|
|---|
| Project Period (FY) |
1988 – 1990
|
| Keywords | Plural information / Image information / Linguistic information / Natural language processing / Image processing / Meaning representation / 04overall comprehension of image - linguistic information |
|---|
| Research Abstract |
1. Construction of Plural Information Processing System
We have constructed a system on the workstation (SUN-4), which is able to handle both image and linguistic information, based on basic examination which we also have done. This system consists of two parts ; linguistic processing part and visual processing part. Our results of each part and integrated system are described below.
(a) Linguistic Processing Part
We have arranged representations of natural language ; which refer to a shape and episodes about that. And we constructed the language processing system which analyzes natural language sentences and extracts meanings of those ; based on our work of meaning representation methods fit to handling connection of both image and linguistic information.
(b) Image Processing Part
The arrangements of features that we should extract for the connection of visual and linguistic information has done. And we constructed the system which extracts features needed, refering to the model of the object ; 3D singular object.
(c) Integrated System
Integrating the systems above, we constructed the 3D object data base system which owned natural language interface. This data base holds image data of 3D singular objects and episodes of those. In this system, image processing part extracts features such as that of shape of object at registratering data, then the data is registrated in the same way of the episodes given by natural language does. Through this methods above, we have realized the system which is able to handle both image and linguistic information in the same way.
2. Summary and Arrangement of problems in the future.
Under the knowledge got through our research described above, we summarized overall comprehension of plural information at this step, and arranged the problems we should solve from the view of both linguistic processing, image processing, and integrated system sides.
|
