基于卷积神经网络的交通声音事件识别方法Traffic sound event recognition method based on convolutional neural network
张文涛,韩莹莹,黎恒
摘要(Abstract)：
针对公路交通声音事件识别中传统语音算法识别效率低、鲁棒性差的问题,提出一种基于卷积神经网络的交通声音事件识别方法。首先通过Gammatone滤波器对声音数字信号进行子带滤波,得到音频信号耳蜗谱图,然后将其代入卷积神经网络模型对声音事件类型进行识别。利用上述方法对公路交通环境下的四种音频事件做了检测处理,并与经典的隐马尔科夫模型和目前广泛使用的深层神经网络进行比较。实验结果表明,使用卷积神经网络模型能够更加准确地对交通声音事件进行识别,且在噪声环境下具有更好的鲁棒性。
关键词(KeyWords)： Gammatone滤波器;卷积神经网络;音频事件识别;公路交通环境;声音数字信号;子带滤波
基金项目(Foundation): 国家自然科学基金(61565004);; 广西自然科学基金(2014GXNSFGA118003);; 桂林市科学研究与技术开发课题(20140127-1;20150133-3)~~
作者(Author): 张文涛,韩莹莹,黎恒
DOI: 10.16652/j.issn.1004-373x.2018.14.018
参考文献(References)：
- [1]MCLOUGHLIN I,ZHANG H,XIE Z,et al.Robust sound event classification using deep neural networks[J].IEEE/ACM transactions on audio,speech,and language processing,2015,23(3):540-552.
- [2]KUCUKBAY S E,SERT M.Audio-based event detection in office live environments using optimized MFCC-SVM approach[C]//Proceedings of IEEE International Conference on Semantic Computing.Anaheim:IEEE,2015:475-480.
- [3]DIMENT A,HEITTOLA T,VIRTANEN T.Sound event detection for office live and office synthetic AASP challenge[J/OL].[2013-12-01].http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.706.807&rep=rep1&type=pdf.
- [4]ABDEL-HAMID O,MOHAMED A,JIANG H,et al.Applying convolutional neural networks concepts to hybrid NN-HMM model for speech recognition[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing.Kyoto:IEEE,2012:4277-4280.
- [5]LU B,DIBAZAR A,BERGER T W.Noise-robust acoustic signature recognition using nonlinear Hebbian learning[J].Neural networks,2010,23(10):1252-1263.
- [6]TJANDRA A,SAKTI S,NEUBIG G,et al.Combination of twodimensional cochleogram and spectrogram features for deep learning-based ASR[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing.Brisbane:IEEE,2015:4525-4529.
- [7]LECUN Y,BOSER B,DENKER J S,et al.Backpropagation applied to handwritten zip code recognition[J].Neural computation,1989,1(4):541-551.
- [8]汪济民,陆建峰.基于卷积神经网络的人脸性别识别[J].现代电子技术,2015,38(7):81-84.WANG Jimin,LU Jianfeng.Face gender recognition based on convolutional neural network[J].Modern electronics technique,2015,38(7):81-84.
- [9]JI S,XU W,YANG M,et al.3D convolutional neural networks for human action recognition[J].IEEE transactions on pattern analysis and machine intelligence,2013,35(1):221-231.
- [10]黄琳,张尤赛.应用深层卷积神经网络的交通标志识别[J].现代电子技术,2015,38(13):101-106.HUANG Lin,ZHANG Yousai.Traffic signs recognition applying with deep-layer convolution neural network[J].Modern electronics technique,2015,38(13):101-106.
- [11]TEMKO A,NADEU C,MACHO D,et al.Acoustic event de-tection and classification[M]//WAIBEL A,STIEFELHAGENR.Computers in the human interaction loop.Berlin:Spring-er,2009:61-73.
- [12]TEMKO A,MALKIN R,ZIEGER C,et al.CLEAR evalua-tion of acoustic event detection and classification systems[C]//Proceedings of the 1st international evaluation conference onclassification of events,activities and relationships.Berlin:Springer,2006:311-322.