【摘要】 研究了基于两步法的欠定语音信号盲分离。针对混合信号散点图在原点中心混叠程度过高的缺点,提出了弭灭圆K均值聚类算法,提高了混叠矩阵的估计精度。结合时频分析算法实现了欠定瞬时线性混叠语音信号的盲分离,取得了较好的分离效果。更多还原

【Abstract】 Sparse component analysis is a signal processing method based on sparse representation.In order to overcome the shortcoming that the aliasing level of the mixed-signal scatter at the origin center is too high,based on sparse component analysis method,a new K means clustering algorithm which can estimate mixing matrix more effectively is proposed.Combined with time-frequency analysis,the proposed algorithm can achieve the instantaneous linear aliasing blind separation of speech signals and get a good separation effect.更多还原