Symbolic pattern retrieval using onomatopoeic voice queries including out-of-vocabulary words for automatic speech recognizers
Project/Area Number |
24500131
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Media informatics/Database
|
Research Institution | Tokyo University of Technology |
Principal Investigator |
AIKAWA Kiyoaki 東京工科大学, メディア学部, 教授 (00367202)
|
Project Period (FY) |
2012-04-01 – 2015-03-31
|
Project Status |
Completed (Fiscal Year 2014)
|
Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2014: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2013: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2012: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
|
Keywords | 音声検索 / 擬態語 / 擬音語 / 未知語 / 感性情報 / パタン検索 / 音響特徴 / 音声認識 / 情報検索 |
Outline of Final Research Achievements |
The objective of this research was the development of voice search methods for items which have no fixed linguistic expressions. Two methods were developed. One was the method using Out-Of-Vocabulary processing using phone sequences obtained by speech recognition. The other was the method using super-segmental acoustic features which were not used in speech recognition. The methods were evaluated on the retrieval "STAMPs" for "LINE" communication. Decision-based methods and similarity measure-based method were compared. The latter method outperformed the former method. The search method was also evaluated on sound-effect retrieval using onomatopoeic voice input.
|
Report
(4 results)
Research Products
(8 results)