STUDIES ON GRAPHIC PRIMITIVES REPRESENTATION OF NATURAL IMAGES BY HIERARCHICAL PROCESSING
| Project/Area Number |
61460128
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
電子通信系統工学
|
|---|
| Research Institution | UNIVERSITY OF TOKYO |
|---|
Principal Investigator |
YASUDA YASUHIKO PROFESSOR INSTITUTE OF INDUSTRIAL SCIENCE UNIVERSITY OF TOKYO, 生産技術研究所, 教授 (40013106)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KATO SHIGEO RESEARCH ASSOCIATE INSTITUTE OF INDUSTRIAL SCIENCE UNIVERSITY OF TOKYO, 生産技術研究所, 助手 (00143529)
ISHIZUKA MITSURU ASSOCIATE PROFESSOR INSTITUTE OF INSUSTRIAL SCIENCE UNIVERSITY OF TOKYO (BABA,Hiroshi), 生産技術研究所, 助教授 (50114369)
|
|---|
| Project Period (FY) |
1986 – 1987
|
| Project Status |
Completed (Fiscal Year 1987)
|
| Budget Amount *help |
¥6,600,000 (Direct Cost: ¥6,600,000)
Fiscal Year 1987: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 1986: ¥6,100,000 (Direct Cost: ¥6,100,000)
|
|---|
| Keywords | Image Coding / Graphic Primitives / Image Command / 階層構成 / 自然画像 / 階層的構造化 / 知的画像符号化 / 画像データベース |
|---|
| Research Abstract |
It is of vital importance to transmit. storage and display effectively in the image information service systems using communication lines.
We proposed a new method usitable to these systems, by representing natural images by graphic primitives such as points, lines and polygons.
In our method. an input natural color image is first converted into a multi-color image using the vector quantization method. Then. the multi-color image is decomposed into region images, with regions to be converted into the command called object regions. Object regions are arranged from larger to smaller. Finally, the ordered object regions are represented by the graphic primitives of the CCITT Data Syntax III. The essence of our method is as follows.
It is desirable for effective command coding to use the inclusion relation among object regions. It is, however, difficult to describe the inclusion relation in multi-color images, whereas. it's easy in two-color images. The multi-color image is, therefore. divided hierarchically into region image. Object regions are expressed by only two colors, according to a binary tree representation, in which each node is classified by its value, i.e., a one further divided into sub-regions, or a terminal one. The hierarchical process is performed from bottom to top.
We only have to consider just two colors when coding layer by this process. The regions extracted by the above mentioned method are converted into the INCREMENTAL POLYGON command of NAPLPS by automatically tracing the borders of regions.
Furthermore, we studied the representation of moving objects by the frame-data structure as an approach to knowledge-based image coding.
|
Report
(2 results)
Research Products
(14 results)
