|
JAIST Repository >
b. 情報科学研究科・情報科学系 >
b11. 会議発表論文・発表資料等 >
b11-1. 会議発表論文・発表資料 >
このアイテムの引用には次の識別子を使用してください:
http://hdl.handle.net/10119/10730
|
| タイトル: | Privacy protection for speech based on concepts of auditory scene analysis |
| 著者: | AKAGI, Masato
IRIE, Yoshihiro |
| 発行日: | 2012-08-22 |
| 出版者: | Institute of Noise Control Engineering of the USA |
| 誌名: | Proc. INTERNOISE 2012 |
| 開始ページ: | 485 |
| 抄録: | There are many sounds in the environment in which we live. In order to obtain the relevant information from a particular sound, we need to identify a target sound and extract the information from that. Being able to pinpoint a target sound from among two or more sounds is called the “cocktail party effect”. The main component of this effect is auditory scene analysis (ASA), which, if it is working well, enables us to focus on something as soft as a whisper uttered several meters ahead. However, there is also the possibility that private conversations—e.g., those that take place at a bank, hospital outpatient department, pharmacy, etc.— may be overheard by another person. In this paper, we propose a method for protecting speech based on circumventing ASA. The proposed method renders the content of an utterance indistinct by generating an acoustical environment in which ASA does not work effectively. It does this by using an information masking technique to reproduce a sound that obscures the phonemes bearing the language semantic information of the conversations. |
| Rights: | Akagi, M. and Irie, Y. (2012/08/22). “Privacy protection for speech based on concepts of auditory scene analysis,” Proc. INTERNOISE2012, New York, 485. |
| URI: | http://hdl.handle.net/10119/10730 |
| 資料タイプ: | publisher |
| 出現コレクション: | b11-1. 会議発表論文・発表資料 (Conference Papers)
|
このアイテムのファイル:
| ファイル |
記述 |
サイズ | 形式 |
|---|
| internoise2012_Akagi.pdf | | 567Kb | Adobe PDF | 見る/開く |
|
当システムに保管されているアイテムはすべて著作権により保護されています。
|
