摘要
为了提高不同背景噪声下语音活动检测的准确性,提出了一种基于背景噪声分类的语音活动检测算法,将常见的环境噪声进行分类,并提出一种子带能量比特征,利用云模型快速准确的识别噪声,从而将开放环境下的检测问题转化成限定环境下的检测问题。实验结果表明,该算法通过背景噪声识别,动态的选择合适的声学特征及相关参数进行语音活动检测,可以有效地降低背景噪声变化带来的影响,提高检测正确率。
In order to improve the accuracy of voice activity detection(VAD) under different background noise, the VAD based on background noise classification is proposed in this paper. Common environmental noise is classified and a feature of sub-band energy ratio id proposed in this paper. Then the method identifies noise quickly and accurately using cloud model, which transform the detection in the open environment into the limited environment. Experimental results show that the proposed VAD by identifying background noise and choosing proper acoustic features dynamically can effectively reduce the effects of background noise and improve the accuracy.
引文
[1]Beritelli F,Casale S,Ruggeri G.Performance evaluation and comparison of ITU-T/ETSI voice activity detectors[C]//Acoustics,Speech,&Signal Processing,on IEEE International Conference.IEEE,2001:1425-1428 vol.3.
[2]Srinivasan K,Gersho A.Voice activity detection for cellular networks[C]//Speech Coding for Telecommunications,1993.Proceedings.IEEE Workshop on.1993:85-86.
[3]Karray L,Martin A.Towards improving speech detection robustness for speech recognition in adverse conditions[J].Speech Communication,2003,40(3):261-276.[4][Woo K H,Yang T Y,Park K J,et al.Robust voice activity detection algorithm for estimating noise spectrum[J].Electronics Letters,2000,36(2):180-181.
[5]舒倩,李银国.基于MFCC0的语音端点检测方法[J].通信技术,2007,40(11):374-375.
[6]朱晓晶,侯旭初,崔慧娟,等.基于LPCC和能量熵的端点检测[J].电讯技术,2010,50(6):41-45.
[7]Saeedi J,Ahadi S M,Faez K.Robust voice activity detection directed by noise classification[J].Signal Image&Video Processing,2013,9(3):1-13.
[8]李祖鹏,姚佩阳.一种语音段起止端点检测新方法[J].电讯技术,2000,40(3):68-70.
[9]李德毅,刘常昱,杜鹢,等.不确定性人工智能[M].国防工业出版社,2014.[10][De L I.Mining Association Rules with Linguistic Cloud Models[J].Journal of Software,2000,1394:0045.
[11]Varga A,Steeneken H J M.Assessment for automatic speech recognition:II.NOISEX-92:A database and an experiment to study the effect of additive noise on speech recognition systems.Speech Communication,1993;12(93):247–251
[12]Canny J.A Computational Approach to Edge Detection[J].Pattern Analysis&Machine Intelligence IEEE Transactions on,1986,PAMI-8(6):679-698.
[13]Petrou M,Kittler J.Optimal edge detectors for ramp edges[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,1991,13(5):483-491.
[14]陈振标,徐波.基于子带能量特征的最优化语音端点检测算法研究[J].声学学报:中文版,2005(2):171-176.
