多噪声环境下的层级语音识别模型

doi:10.11772/j.issn.1001-9081.2017112678

计算机应用 ›› 2018, Vol. 38 ›› Issue (6): 1790-1794.DOI: 10.11772/j.issn.1001-9081.2017112678

• 虚拟现实与多媒体计算 • 上一篇下一篇

多噪声环境下的层级语音识别模型

曹晶晶, 许洁萍, 邵聖淇

中国人民大学信息学院, 北京 100872

收稿日期:2017-11-14 修回日期:2018-01-09 出版日期:2018-06-10 发布日期:2018-06-13
通讯作者: 许洁萍
作者简介:曹晶晶(1993-),女,安徽马鞍山人,硕士研究生,主要研究方向:语音识别;许洁萍(1966-),女,黑龙江牡丹江人,副教授,博士,CCF会员,主要研究方向:音频信息处理;邵聖淇(1993-),男,辽宁沈阳人,硕士研究生,主要研究方向:语音识别。
基金资助:
国家自然科学基金资助项目（61672523）。

Hierarchical speech recognition model in multi-noise environment

CAO Jingjing, XU Jieping, SHAO Shengqi

School of Information, Renmin University of China, Beijing 100872, China

Received:2017-11-14 Revised:2018-01-09 Online:2018-06-10 Published:2018-06-13
Supported by:
This work is partially supported by the National Natural Science Foundation of China (61672523).

摘要/Abstract

摘要： 针对多噪声环境下的语音识别问题，提出了将环境噪声作为语音识别上下文考虑的层级语音识别模型。该模型由含噪语音分类模型和特定噪声环境下的声学模型两层组成，通过含噪语音分类模型降低训练数据与测试数据的差异，消除了特征空间研究对噪声稳定性的限制，并且克服了传统多类型训练在某些噪声环境下识别准确率低的弊端，又通过深度神经网络（DNN）进行声学模型建模，进一步增强声学模型分辨噪声的能力，从而提高模型空间语音识别的噪声鲁棒性。实验中将所提模型与多类型训练得到的基准模型进行对比，结果显示所提层级语音识别模型较该基准模型的词错率（WER）相对降低了20.3%，表明该层级语音识别模型有利于增强语音识别的噪声鲁棒性。

关键词: 语音识别, 噪声鲁棒性, 环境噪声, 声学模型, 深度神经网络

Abstract: Focusing on the issue of speech recognition in multi-noise environment, a new hierarchical speech recognition model considering environmental noise as the context of speech recognition was proposed. The proposed model was composed of two layers of noisy speech classification model and acoustic model under specific noise environment. The difference between training data and test data was reduced by noisy speech classification model, which eliminated the limitation of noise stability required in feature space research and solved the disadvantage of low recognition rate caused by traditional multi-type training under certain noise environment. Furthermore, a Deep Neural Network (DNN) was used for modeling of acoustic model, which could further enhance the ability of acoustic model to distinguish noise and speech, and the noise robustness of speech recognition in model space was improved. In the experiment, the proposed model was compared with the benchmark model obtained by multi-type training. The experimental results show that, the proposed hierarchical speech recognition model has relatively reduced the Word Error Rate (WER) by 20.3% compared with the traditional benchmark model. The proposed hierarchical speech recognition model is helpful to enhance the noise robustness of speech recognition.

Key words: speech recognition, noise-robustness, environmental noise, acoustic model, Deep Neural Network (DNN)

中图分类号:

TP391.4

曹晶晶, 许洁萍, 邵聖淇. 多噪声环境下的层级语音识别模型[J]. 计算机应用, 2018, 38(6): 1790-1794.

CAO Jingjing, XU Jieping, SHAO Shengqi. Hierarchical speech recognition model in multi-noise environment[J]. Journal of Computer Applications, 2018, 38(6): 1790-1794.

参考文献

[1] LI J Y, DENG L, GONG Y F, et al. An overview of noise-robust automatic speech recognition[J]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2014, 22(4):745-777.
[2] HIMAWAN I, MOTLICEK P, IMSENG D, et al. Learning feature mapping using deep neural network bottleneck features for distant large vocabulary speech recognition[C]//Proceedings of the 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing. Piscataway, NJ:IEEE, 2015:4540-4544.
[3] HAN K, HE Y Z, BAGCHI D, et al. Deep neural network based spectral feature mapping for robust speech recognition[C]//Proceedings of the 201516th Annual Conference of the International Speech Communication Association. Grenoble, France:ISCA, 2015:2484-2488.
[4] REHR R, GERKMANN T. Cepstral noise subtraction for robust automatic speech recognition[C]//Proceedings of the 2015 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway, NJ:IEEE, 2015:375-378.
[5] WANG D, ZHANG X W. THCHS-30:a free Chinese speech corpus[EB/OL].[2017-10-16]. http://pdfs.semanticscholar.org/207e/c1b9457c1e42f34d331cf2a7bc791358b9cd.pdf.
[6] LIPPMANN R, MARTIN E, PAUL D. Multi-style training for robust isolated-word speech recognition[C]//Proceedings of the 2003 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway, NJ:IEEE, 2003:705-708.
[7] 易克初,田斌,付强.语音信号处理[M].北京:国防工业出版社,2000:210-242.(YI K C, TIAN B, FU Q. Speech Signal Processing[M]. Beijing:National Defense Industry Press, 2000:210-242.)
[8] 张仕良.基于深度神经网络的语音识别模型研究[D].合肥:中国科学技术大学,2017:1-4.(ZHANG S L. Research on deep neural network based models for speech recognition[D]. Hefei:University of Science and Technology of China, 2017:1-4.)
[9] DAHL G E, YU D, DENG L, et al. Context-dependent pre-trained deep neural networks for large-vocabulary speech recognition[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2012, 20(1):30-42.
[10] GAO T, DU J, DAI L R, et al. Joint training of front-end and back-end deep neural networks for robust speech recognition[C]//Proceedings of the 2015 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway, NJ:IEEE, 2015:4375-4379.
[11] MA L, MILNER B, SMITH D. Acoustic environment classification[J]. ACM Transactions on Speech and Language Processing, 2006, 3(2):1-22.
[12] CHU S, NARAYANAN S, KUO C C J. Environmental sound recognition with time-frequency audio features[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2009, 17(6):1142-1158.
[13] XUE X B, ZHOU Z H. Distributional features for text categorization[J]. IEEE Transactions on Knowledge and Data Engineering, 2009, 21(3):428-442.
[14] 周志华.机器学习[M].北京:清华大学出版社,2016:121-145.(ZHOU Z H. Machine Learning[M]. Beijing:Tsinghua University Press, 2016:121-145.)
[15] PHILBIN J, CHUM O, ISARD M, et al. Object retrieval with large vocabularies and fast spatial matching[C]//Proceedings of the 2007 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ:IEEE, 2007:1-8.
[16] LIANG J W, JIN Q, HE X X, et al. Semantic concept annotation of consumer videos at frame-level using audio[C]//Proceedings of the 201415th Pacific-Rim Conference on Advances in Multimedia Information Processing, LNCS 8879. Cham:Springer, 2014:113-122.
[17] VESELY K, GHOSHAL A, BURGET L, et al. Sequence-discriminative training of deep neural networks[C]//Proceedings of the 201314th Annual Conference of International Speech Communication Association. Prefecture of Grenoble, France:ISCA, 2013:2345-2349.
[18] 俞栋,邓力.解析深度学习:语音识别实践[M].北京:电子工业出版社,2016:81-85.(YU D, DENG L. Parsing the Deep Learning:Speech Recognition Practices[M]. Beijing:Publishing House of Electronics Industry, 2016:81-85.)

多噪声环境下的层级语音识别模型

Hierarchical speech recognition model in multi-noise environment

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	王曙燕, 侯则昱, 孙家泽. 面向深度学习的对抗样本差异性检测方法[J]. 计算机应用, 2021, 41(7): 1849-1856.
[2]	张明明, 卢庆宁, 李文中, 宋浒. 基于联合动态剪枝的深度神经网络压缩算法[J]. 计算机应用, 2021, 41(6): 1589-1596.
[3]	张文烨, 尚方信, 郭浩. 基于Octave卷积的混合精度神经网络量化方法[J]. 计算机应用, 2021, 41(5): 1299-1304.
[4]	杨丽, 王时绘, 朱博. 基于动态和静态偏好的兴趣点推荐算法[J]. 计算机应用, 2021, 41(2): 398-406.
[5]	陈彦如, 张涂静娃, 杜千, 冉茂亮, 王红军. 基于深度森林的高铁站室内热舒适度等级预测[J]. 计算机应用, 2021, 41(1): 258-264.
[6]	李鸣, 郭晨皓, 陈星. 视觉类深度神经网络的自动标注[J]. 计算机应用, 2020, 40(6): 1593-1600.
[7]	陈玉娜, 史晓东. 通过标点恢复提高机器同传效果[J]. 计算机应用, 2020, 40(4): 972-977.
[8]	韦伟, 李小娟. 基于相似论文增广的深度学习专利质量评估[J]. 计算机应用, 2020, 40(4): 966-971.
[9]	邓凯, 黄佳进, 秦进. 基于物品的统一推荐模型[J]. 计算机应用, 2020, 40(2): 530-534.
[10]	樊琦, 李卓, 陈昕. 基于边缘计算的分支神经网络模型推断延迟优化[J]. 计算机应用, 2020, 40(2): 342-346.
[11]	杨坚伟, 严群, 姚剑敏, 林志贤. 基于深度神经网络的移动端人像分割[J]. 计算机应用, 2020, 40(12): 3644-3650.
[12]	杨磊, 赵红东. 基于轻量级深度神经网络的环境声音识别[J]. 计算机应用, 2020, 40(11): 3172-3177.
[13]	朱倩倩, 刘渊, 李甫. 深度神经网络的仿生矩阵约简与量化方法[J]. 计算机应用, 2020, 40(10): 2817-2821.
[14]	严红, 陈兴蜀, 王文贤, 王海舟, 殷明勇. 基于深度神经网络的法语命名实体识别模型[J]. 计算机应用, 2019, 39(5): 1288-1292.
[15]	陈龙杰, 张钰, 张玉梅, 吴晓军. 基于多注意力多尺度特征融合的图像描述生成算法[J]. 计算机应用, 2019, 39(2): 354-359.