节点文献
用通道补偿提高多通道语音系统识别率的新方法
TO IMPROVE THE RECOGNITION RATE OF MICROPHONE-ARRAY SPEECH SYSTEMS BY CHANNEL-COMPENSATION
【摘要】 本文提出一种针对多通道语音系统的通道补偿方法来提高系统对语音通道变化的识别鲁棒性。它利用噪声子空间和信号子空间的正交性对通道系数进行识别,进而构造均衡器恢复出源信号。该算法在两个通道的情况下就能收到很好的效果。它克服了现有的多通道语音处理方法中没有考虑通道影响的问题,并且也比单通道条件下的通道补偿方法有更好的效果。
【Abstract】 A channel-compensation method for microphone-array systems is proposed. It utilizes the orthogonal property of the signal subspace and the noise subspace to identify the channel coefficients. A two-microphone system is sufficient for the algorithm to work well. The computer simulation shows thai it can improve the robustness of speech/speaker recognition systems by reducing the effect of channel and reverberance.
【关键词】 多通道语音系统;
语音通道补偿;
子空间分解;
【Key words】 Microphone-Array Systems; Speech Channel Compensation; Suhspace Decomposition;
【Key words】 Microphone-Array Systems; Speech Channel Compensation; Suhspace Decomposition;
- 【文献出处】 模式识别与人工智能 ,Pattern Recognition and Artificial Intelligence , 编辑部邮箱 ,2003年03期
- 【分类号】TN912.34
- 【下载频次】48