摘要
本文基于盲定位(Blind Source Localization,BSL)指导盲分离(Blind Source Separation,BSS)的思想,提出了一种用无限冲激响应(IIR)滤波器来构造前馈分离网络的盲分离算法,该算法不仅解决了传统盲分离算法不能估计信号源数目和不能确定分离输出信号与源信号之间的对应关系的问题,还能在实施信号分离处理之前判断所构造前馈分离网络的稳定性,对语音信号的仿真实验验证了该算法的有效性。
An algorithm for the purpose of sound source separation is presented in this paper. The algorithm is based on the inherent relationship between blind source separation (BSS) and blind source localization (BSL). The correponding feedforward network for separating sound sources can be realized by IIR filters. The algorithm can solve both the problem of estimating the number of sound sources and the problem of determining the correspondence between outputs and inputs. It can also give some results about the stability of the network before the system starts the work of separation. The experimental results show the effciency of the algorithm.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2001年第3期285-290,共6页
Pattern Recognition and Artificial Intelligence
