location: Current position: Home >> Scientific Research >> Paper Publications

基于分层采样粒子滤波的麦克风阵列说话人跟踪方法

Hits:

Indexed by:期刊论文

Date of Publication:2008-01-15

Journal:电子学报

Included Journals:EI、PKU、ISTIC、CSCD

Volume:36

Issue:1

Page Number:194-198

ISSN No.:0372-2112

Key Words:说话人跟踪;麦克风阵列;粒子滤波;分层采样;房间混响

Abstract:针对噪声与混响环境下的说话人跟踪问题,本文提出了一种基于粒子滤波的麦克风阵列声源定位与跟踪方法.该方法在粒子滤波框架下,将无混响影响的语音建立信号作为观测信息,通过计算麦克风阵列波束形成器的输出能量来构建似然函数,同时考虑语音信号不同频率成分在声源定位中的作用,利用分层采样方法提高粒子的采样效率.实验结果表明,本文方法提高了说话人声源跟踪系统的抗噪声与抗混响能力.

Pre One:HRTF compression via principal components analysis and vector quantization

Next One:Optimum design of a window function based on the small-world networks