基于改进型BP神经网络的音频多分类
|
摘要
音频信号作为多媒体信息的重要载体之一,为满足人们对信息知识的获取提供了有效途径.为了提高音频分类的精度,提出一种将音频信号的梅尔频率倒谱系数(Mel frequency cepstrum coefficient,MFCC)参数作为特征向量,采用基于改进型传输函数的误差反向传播神经(back propagation,BP)网络模型对6种音频进行分类.实验证明,该方法在音频分类精度方面性能良好,改进的传输函数具有收敛速度快的优点.相对于传统BP算法,该方法不仅缩短了训练时间,而且进一步提高了分类精度,其分类准确率达到90%以上.
Audio is an important medium that carries substantial information to meet human needs.To improve accuracy of audio classification,we propose a new algorithm with Mel frequency cepstrum coefficient(MFCC) parameters as the feature vectors,and use a back propagation(BP) neural network model based on improved transfer function to classify six types of audio signals.Experiments show that the proposed algorithm has good performance and the improved transfer function converges faster that the traditional BP algorithm.It can reduce training time,and improve classification accuracy up to more than 90%.
Audio is an important medium that carries substantial information to meet human needs.To improve accuracy of audio classification,we propose a new algorithm with Mel frequency cepstrum coefficient(MFCC) parameters as the feature vectors,and use a back propagation(BP) neural network model based on improved transfer function to classify six types of audio signals.Experiments show that the proposed algorithm has good performance and the improved transfer function converges faster that the traditional BP algorithm.It can reduce training time,and improve classification accuracy up to more than 90%.
引文
[1]WOLD E,BLUM T,KEISLAR D,et al.Content-basedclassification,search and retrieval of audio[J].IEEEMultimedia,1996,3(3):27-36.
[2]丁爱明.作为说话人识别特征参量的MFCC的提取过程[J].电子工程师,2006,32(1):51-53.
[3]SCHEIRER E,SLANEY M.Construction and evaluationof a robust multi-feature speech/music discriminator[C]∥Proceedings of the 1997 IEEE International Conferenceon Acoustics,Speech,and Signal Processing.1997:1331-1334.
[4]周志华,曹存根.神经网络及其应用[M].北京:清华大学出版社,2004:12-33.
[5]MENG Z P,TIAN Y D,LEI Y.Prediction models of coalbed gas content based on BP neural networks and itsapplications[J].Journal of China University of Mining&Technology,2008,37(4):456-461.
[6]LI K,HUANG K H.A selective neural networkintegration based on aggressive classes technique[J].Journal of Computer Research and Development,2007,42(4):594-598.
[7]王华,程海清.自适应动量项BP神经网络盲均衡算法[J].计算机工程与设计,2010,31(6):1297-1300.
[8]WU W,XU Y S.Deterministic convergence of an onlinegradient method for neural networks[J].Journal ofComputational and Applied Mathematics,2002,144(1):335-347.
[9]蔡永香,郭庆胜,桂志先,等.基于地震-测井数据预测储层参数空间分布规律的神经网络模型[J].武汉大学学报:信息科学版,2005,30(4):366-369.
[10]XU F,LIU F,GU W J,et al.A clustering algorithmbased on constructive neural network with feedbackconnections[J].Computer Engineer and Application,2004,40(20):50-53.
[11]TZANETAKIS G,COOK P.Musical genre classification ofaudio signals[J].IEEE Transactions on Speech andAudio Processing,2002,10(5):293-302.
[2]丁爱明.作为说话人识别特征参量的MFCC的提取过程[J].电子工程师,2006,32(1):51-53.
[3]SCHEIRER E,SLANEY M.Construction and evaluationof a robust multi-feature speech/music discriminator[C]∥Proceedings of the 1997 IEEE International Conferenceon Acoustics,Speech,and Signal Processing.1997:1331-1334.
[4]周志华,曹存根.神经网络及其应用[M].北京:清华大学出版社,2004:12-33.
[5]MENG Z P,TIAN Y D,LEI Y.Prediction models of coalbed gas content based on BP neural networks and itsapplications[J].Journal of China University of Mining&Technology,2008,37(4):456-461.
[6]LI K,HUANG K H.A selective neural networkintegration based on aggressive classes technique[J].Journal of Computer Research and Development,2007,42(4):594-598.
[7]王华,程海清.自适应动量项BP神经网络盲均衡算法[J].计算机工程与设计,2010,31(6):1297-1300.
[8]WU W,XU Y S.Deterministic convergence of an onlinegradient method for neural networks[J].Journal ofComputational and Applied Mathematics,2002,144(1):335-347.
[9]蔡永香,郭庆胜,桂志先,等.基于地震-测井数据预测储层参数空间分布规律的神经网络模型[J].武汉大学学报:信息科学版,2005,30(4):366-369.
[10]XU F,LIU F,GU W J,et al.A clustering algorithmbased on constructive neural network with feedbackconnections[J].Computer Engineer and Application,2004,40(20):50-53.
[11]TZANETAKIS G,COOK P.Musical genre classification ofaudio signals[J].IEEE Transactions on Speech andAudio Processing,2002,10(5):293-302.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心