| Project/Area Number |
11650385
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | KYUSHU UNIVERSITY |
|---|
Principal Investigator |
KAMATA Sei-ichiro Faculty of Information Science and Electrical Engineering, KYUSHU UNIVERSITY, Ass.Prof., 大学院・システム情報科学研究院, 助教授 (00204602)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KATAYAMA Yoshinori Faculty of Information Science and Electrical Engineering, KYUSHU UNIVERSITY, Ass.Researcher, 大学院・システム情報科学研究院, 助手 (00214338)
SAKOE Hiroaki Faculty of Information Science and Electrical Engineering, KYUSHU UNIVERSITY, Prof., 大学院・システム情報科学研究院, 教授 (30215701)
|
|---|
| Project Period (FY) |
1999 – 2000
|
| Project Status |
Completed (Fiscal Year 2000)
|
| Budget Amount *help |
¥3,200,000 (Direct Cost: ¥3,200,000)
Fiscal Year 2000: ¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 1999: ¥2,000,000 (Direct Cost: ¥2,000,000)
|
|---|
| Keywords | space-filling scan / Peano curve / image coding / image display / image processing / image retrieval / image representation / 面像符号化 |
|---|
| Research Abstract |
Techniques for not only compression, but also search engine and content extraction of image database are essential for next generation of multimedia information processing such as MPEG7. The purpose of this research is to develop a compression technique for searching images in a database using a space-filling curve. The obtained research achievements are as follows. (1) First, we realized a scan using a space-filling curve which pass through all points for each region segmented in an image. This method is based on building a Minimum Spanning Tree (MST) from the image and traversing the MST.We also developed this algorithm to improve compression efficiency using a cutting technique of MST.(2) Next, we improved compression efficiency using trellis coded vector quantization (TCVQ). TCVQ is an application of trellis code modulation which is a well known technique in wave transmition area. Using some real images, the SN ratio improved 0.5 〜 1.0 dB for the same compression rate. (3) Finally, using gray images and color images, the relation between compression efficiency of our method and comptation complexity is investigated in comparison with the conventional methods. From several experimental results, the quality of compressed images are comparable to the one of JPEG.However, our method is about 20 times slower than JPEG.This problem needs to be solved in future.
|
