open access

https://hdl.handle.net/2115/39778

Component reduction technique for covariance matrix of multidimensional Gaussian distribution in Speech Recognition

英語

2009-10-04

Asia-Pacific Signal and Information Processing Association, 2009 Annual Summit and Conference, International Organizing Committee

Proceedings : APSIPA ASC 2009 : Asia-Pacific Signal and Information Processing Association, 2009 Annual Summit and Conference

644-647

APSIPA ASC 2009: Asia-Pacific Signal and Information Processing Association, 2009 Annual Summit and Conference. 4-7 October 2009. Sapporo, Japan. Poster session: Automatic Speech Recognition (6 October 2009).

Recently, speech recognition systems are in practical use as an input device of a car navigation system, etc. However, since the computational and storage resources are usually limited for a car navigation system, these systems are required to achieve high recognition performance with the limited resource. In this paper, we propose a method that reduces the components of full covariance matrix considering only the dominant correlation components between static and dynamic feature parameters. The recognition experiment was performed by using the conventional and the proposed method, and both were compared. From the continuous syllable recognition result, we confirmed the effectiveness of proposed component reduction technique considering the correlation between parameters.

会議発表論文

Version of Record