|
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/10119/9978
|
| タイトル: | Towards an intelligent binaural speech enhancement system by integrating meaningful signal extraction |
| 著者: | Chau, Duc Thanh
Li, Junfeng
Akagi, Masato |
| 発行日: | 2011-03-03 |
| 出版者: | 2011 International Workshop on Nonlinear Circuits, Communication and Signal Processing |
| 誌名: | 2011 International Workshop on Nonlinear Circuits, Communication and Signal Processing (NCSP'11) |
| 開始ページ: | 344 |
| 終了ページ: | 347 |
| 抄録: | Current speech enhancement applications, such asbinaural hearing aids, mainly aim to suppress interference signals and enhance the target signal with preservation of binaural cues. However, in addition to the target signal, human beings are able to pay attention to other important or meaningfulsounds (e.g., the call from others) in daily conversation. This attention mechanism to meaningful signals is seldom considered in the state-of-the-art signal processing systems. In this paper, we propose an intelligent binaural speech enhancement model by extracting the meaningful signals as well as the target signal. Specifically, the proposed model consists of two main parallelprocesses: binaural target signal enhancement and binaural meaningful signal extraction, finally yielding the binaural outputs. Experimental result showed that the proposed system is able to not only suppress interfering noise signals, but also enhancetarget signal and meaningful signals. |
| Rights: | This material is posted here with permission of the Research Institute of Signal Processing Japan. Duc Thanh Chau, Junfeng Li and Masato Akagi, 2011 International Workshop on Nonlinear Circuits, Communication and Signal Processing (NCSP'11), 2011, pp.344-347. |
| URI: | http://hdl.handle.net/10119/9978 |
| 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 |
|---|
| NCSP022-Chau.pdf | | 265Kb | 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。
|
