2004 Fiscal Year Final Research Report Summary
Research on Linguistic Understanding of Human Motion for Patient Care Support
| Project/Area Number |
14580436
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Fukuoka Institute of Technology |
|---|
Principal Investigator |
YOKOTA Masao Fukuoka Institute of Technology, Faculty of Information Engineering, Professor, 情報工学部, 教授 (50112313)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ODA Seio Fukuoka Institute of Technology, Junior College, Professor, 教授 (10185598)
|
|---|
| Project Period (FY) |
2002 – 2004
|
| Keywords | Human motion / Motion capturing system / Natural language / Cross-media Translation / Artificial Intelligence |
|---|
| Research Abstract |
Rapid increase of matured societies will necessarily bring a large number of handicapped people due to aging and thereby serious shortage of workers for them such as nurses, one of whose routines is to survey their actions through TV monitors in, order to prevent accidents.
However it is unfeasible and privacy-invasive for human workers to watch TV monitors for such surveillance all day and night long. For such a purpose, certain kinds of intelligent systems can be very helpful that facilitate exploiting significant parts such as abnormal motions from video data and reporting it in natural language in real time and desirably linguistic summarization of immense amount of recorded ones.
M.Yokota, et al have proposed a formal language named L_<md> for multimedia information description in the Mental Image Directed Semantic Theory (MIDST) and thereby a methodology for automatic understanding of various information media through intermediate knowledge representation. The language L_<md> is employed for many-sorted first-order predicate logic containing one special predicate called 'Atomic locus' with five types of terms. The most remarkable feature of L_<md> is its capability of formalizing both temporal and spatial event concepts on the level of human sensations while the other similar knowledge representation languages are designed to describe the logical relations among conceptual primitives such as words.
Employing L_<md>, D.Hironaka et al have been developing an integrated multimedia understanding system IMAGES -M that can facilitate, for example, cross-media reference including bi-directional translation between different kinds of media such as text and picture. M.Amano et al have also been applying L_<md> to systematic linguistic interpretation of human motions.
|
