| Project/Area Number |
08680402
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Kyushu University |
|---|
Principal Investigator |
SAKOE Hirbaki Kyushu University Graduate School of Information Science and Electrical Engineering Professor, 大学院・システム情報科学研究科, 教授 (30215701)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KATAYAMA Yoshnori Kyushu University Graduate School of Information Science and Electrical Engineer, 大学院・システム情報科学研究科, 助手 (00214338)
MIYAZAKI Akio Kyushu University Graduate School of Information Science and Electrical Engineer, 大学院・システム情報科学研究科, 助教授 (70192763)
|
|---|
| Project Period (FY) |
1996 – 1997
|
| Project Status |
Completed (Fiscal Year 1997)
|
| Budget Amount *help |
¥2,100,000 (Direct Cost: ¥2,100,000)
Fiscal Year 1997: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1996: ¥1,300,000 (Direct Cost: ¥1,300,000)
|
|---|
| Keywords | Line Pattern Parsing / Line Pattern Detection / CYK Algorithm / Two Dimensional CYK / Multiple Trajectory parsing / Noise-robustness / Dynamic Programming / Beam Search |
|---|
| Research Abstract |
Algoriths were investigated for analyzing line patterns in noisy images, based on generation model.
(1) So far, several parsing algorithms were reported which analyze two dimensional imagcs, based on a model given in a regular grammar format. The parsing capability of these algorithm were rather low due to the use of regular grammar. In this research. use of context free grammar model were investigated. On CYK algorithm which can parse two dimensional input according to the given CFG model was newly proposed. An improved parsing function and efficiency enhancement were proved through experiments.
(2) An algorithm which parses multiple trajectories in a time based images were investigated. Different from conventional ones, the proposed algorithm parses the input according to the knowledge on mutual relations between trajectories, as well as knowledge on individual trajectory. Noise-robustness were established through experiments.
|
