A study of speech information processing based on mathematical models for speaker and linguistic information and there probabilistic integration

2015 Fiscal Year Final Research Report

• Project/Area Number: 25730105
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Perceptual information processing
• Research Institution: The University of Tokyo
• Principal Investigator: SAITO DAISUKE 東京大学, 情報理工学(系)研究科, 助教 (40615150)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 音声情報処理 / 声質変換 / 話者識別 / 行列変量 / 言語識別 / テンソル解析

Outline of Final Research Achievements

In this study, to achieve more sophisticated speech information processing, mathematical models which divide speech into linguistic information and speaker information separately were developed. In addition, a framework where these mathematical models are integrated was also developed. We have proposed speech representation based on tensor analysis and applied to language identification and speaker identification. A new voice conversion framework based on matrix variate probabilistic distribution was also developed.

Free Research Field

音声情報処理