2005 Fiscal Year Final Research Report Summary
Fuzzy objects based image compression system for managing vast images
| Project/Area Number |
15300069
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Sensitivity informatics/Soft computing
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
HIROTA Kaoru Tokyo Institute of Technology, Interdisciplinary Graduate School of Science and Engineering, Professor, 大学院・総合理工学研究科, 教授 (50130943)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
NOBUHARA Hajime Tokyo Institute of Technology, Interdisciplinary Graduate School of Science and Engineering, Assistant Professor, 大学院・総合理工学研究科, 助手 (80359687)
HATAKEYAMA Yutaka Tokyo Institute of Technology, Interdisciplinary Graduate School of Science and Engineering, Assistant Professor, 大学院・総合理工学研究科, 助手 (00376956)
|
|---|
| Project Period (FY) |
2003 – 2005
|
| Keywords | image compression / reconstruction / soft computing / computational intelligence / fuzzy clustering / wavelets / image database / image restoration / image evaluation |
|---|
| Research Abstract |
A fuzzy objects based image compression system is proposed. As an image compression system based on emerging computational resources in order to manage vast images. The proposed system is composed of four parts, 1. pre-processing part, 2. objects extraction part, 3. fuzzy objects database, and 4 retrieval part.
1. Pre-processing part : Two multi-resolution transformations are proposed based on different mathematical structures. The first one is an image compression and reconstruction method based on fuzzy relational equations (fuzzy algebra), and the second one is morphological wavelets based on max plus algebra.
2. Object extraction part : The original images are decomposed into fuzzy objects by using a fuzzy relation decomposition method proposed in this research. Furthermore, an optical flow based segmentation method is also proposed to decompose the original images.
3. Fuzzy object database part: With respect to the fuzzy objects, an optimized database structure is constructed by using the proposed tree construction algorithm along with pai-t neural networks.
4. Retrieval part : In order to reduce the retrieval time and improve the quality of the reconstructed image obtained by the proposed system, a color instance based retrieval method is proposed.
Through the evaluation experiments using various images (selected from standard image database, Corel gallery, and obtained by surveillance camera), the effectiveness of the proposed image compression system is confirmed.
|
Research Products
(32 results)
