1999 Fiscal Year Final Research Report Summary
Numerical Algorithm for Inverse Problems and its Application to Multimedia Information Systems.
| Project/Area Number |
09650075
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Engineering fundamentals
|
|---|
| Research Institution | University of Tsukuba |
|---|
Principal Investigator |
KITAGAWA Takashi University of Tsukuba Inst. of Information Sciences and Electronics, Associate Professor., 電子・情報工学系, 助教授 (60153095)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SAKURAI Tetsuya University of Tsukuba Inst. of Information Sciences and Electronics, Associate Professor., 電子・情報工学系, 助教授 (60187086)
NATORI Makoto University of Tsukuba Inst. of Information Sciences and Electronics, Professor., 電子・情報工学系, 教授 (70013745)
|
|---|
| Project Period (FY) |
1997 – 1999
|
| Keywords | Inverse Problems / Multimedia Information Systems / Semantic Associative Search / Numerical Algorithms. / 意味的相互運用性 |
|---|
| Research Abstract |
In this research project, we designed and analyzed numerical algorithm for inverse problems. The problems originally comes from inversion of compact operators whose typical example is Fredholm integral equations of the first kind. Our main theme includes developing stable numerical algorithm and giving its mathematical foundations. Moreover we extended our algorithms to multimedia information system and proposed new mathematical model which enables semantic interpretation of data items.
The essential operation of regularization is that selecting an appropriate subspace according to the constraint. We generalized this operation to semantic interpretation where the constraint corresponds to context. We can deal with meaning which is the relation of content and the context using the model. This model is applied to actual information systems including meta-database systems and associative search for multimedia data.
We also developed a learning mechanism to deal with individual differences and an accelerating method to realize real time information retrieval. We actually applied this method to semantic image search, multilingual meta search, media search including music, colours and so on.
|
