A Continuous Speech Recognition Method Using Dependent Feature Transformation and Combination of Subspace Region

doi:10.7652/xjtuxb201604010

您当前的位置：

首页 >

文章列表页 >

A Continuous Speech Recognition Method Using Dependent Feature Transformation and Combination of Subspace Region

更新时间：2025-07-09

- A Continuous Speech Recognition Method Using Dependent Feature Transformation and Combination of Subspace Region
- Vol. 50, Issue 4, Pages: 60-67(2016)
- 作者机构：
  
  解放军信息工程大学信息系统工程学院,郑州,450001
- 作者简介：
- 基金信息：
- DOI：10.7652/xjtuxb201604010
  CLC： TN912
- Online First：10 April 2016，
  
  Published：2016
- 稿件说明：
移动端阅览
A Continuous Speech Recognition Method Using Dependent Feature Transformation and Combination of Subspace Region[J]. 2016, 50(4): 60-67.
DOI：

A Continuous Speech Recognition Method Using Dependent Feature Transformation and Combination of Subspace Region[J]. 2016, 50(4): 60-67. DOI： 10.7652/xjtuxb201604010.

摘要

为了提高语音识别准确率

提出了一种子空间域相关特征变换与融合的语音识别方法(MFCC-BN-TC方法)。该方法提取语音短时谱结构特征(BN)和包络特征(MFCC)分别描述语音短时谱结构和包络信息

并采用域相关特征变换的形式分别对BN和MFCC特征进行特征变换; 然后对这种变换进行泛化扩展提出子空间域相关特征变换

以采用不同的时间颗粒度(帧和语音分段)进行多层次区分性特征表达; 最后

对多种区分性特征变换后的特征进行联合表征训练声学模型

并给出了区分性特征变换与融合的一般框架。实验结果表明:MFCC-BN-TC方法比采用原始BN特征方法和采用MFCC特征基线系统方法

识别性能各自提高了0.98%和1.62%; 融合MFCC-BN-TC方法变换以后的语音信号特征

相比于融合原始特征

识别率提升了1.5%。

Abstract

A speech recognition method based on dependent feature transformation and combination of subspace regions(MFCC-BN-TC)is proposed to improve the recognition accuracy. The structure feature(BN)and envelope feature(MFCC)are extracted to separately describe the structure and envelope information of the short speech spectrum

and the region dependent feature transformation is adopted to perform feature transformation for the BN and the MFCC

respectively.The transformation is then generalized to give a subspace region-dependent feature transformation so that different time units(frame and segment)are applied to finish multi-level modeling. Moreover

a feature combination framework is proposed

and the acoustic model is trained using combined multi-features after transformation. Experimental results and comparisons with the method using raw BN and the method based on MFCC feature show that the recognition rate of the MFCC-BN-TC method increases by 0.96% and 1.62%

respectively. The gain in performance of the MFCC-BN-TC method increases by 1.5% through combining the transformed features.

关键词

Keywords

references

NASERSHARIF B, AKBARI A. SNR-dependent compression of enhanced Mel subband energies for compensation of noise effects on MFCC features [J]. Pattern Recognition Letters, 2011, 28(11): 1320-1326.

刘晓明, 班超帆, 冯晓荣. 失真控制下的短时谱估计语音增强算法 [J]. 西安交通大学学报, 2011, 45(8): 78-84.

LIU Xiaoming, BAN Chaofan, FENG Xiaorong. A short time spectrum estimation algorithm of speech enhancement under the distortion control [J]. Journal of Xi'an Jiaotong University, 2011, 45(8): 78-84.

POVEY D, KINGSBURY B, MANGU L, et al. fMPE: Discriminatively trained features for speech recognition [C]∥Proceedings of the International Conference on Audio, Speech and Signal Processing. Piscataway, NJ, USA: IEEE, 2005: 961-964.

ZHANG B, MATSOUKAS S, SCHWARTZ R. Recent progress on the discriminative region-dependent transform for speech feature extraction [C]∥Proceedings of the Annual Conference of International Speech Communication Association. Baixs, France: ISCA, 2006: 1495-1498.

YAN Z, HUO Q, XU J, et al. Tied-state based discriminative training of context-expanded region-dependent feature transforms for LVCSR [C]∥Proceedings of the International Conference on Audio, Speech and Signal Processing. Piscataway, NJ, USA: IEEE, 2013: 6940-6944.

高莹莹, 朱维彬. 深层神经网络中间层可见化建模 [J]. 自动化学报, 2015, 41(9): 1627-1637.

GAO Yingying, ZHU Weibin. Deep neural networks with visible intermediate layers [J]. Acta Automatica Sinica, 2015, 41(9): 1627-1637.

袁胜龙, 郭武, 戴礼荣. 基于深层神经网络的藏语识别 [J]. 模式识别与人工智能, 2015, 28(3): 209-213.

YUAN Shenglong, GUO Wu, DAI Lirong. Speech recognition based on deep neural networks on Tibetan corpus [J]. Pattern Recognition and Artificial Intelligence, 2015, 28(3): 209-213.

SAINATH T N, KINGSBURY B, RAMABHADRAN B. Auto-encoder bottleneck features using deep belief networks [C]∥Proceedings of the International Conference on Audio, Speech and Signal Processing. Piscataway, NJ, USA: IEEE, 2012: 4153-4156.

SAON G, KINGSBURY B. Discriminative feature-space transforms using deep neural networks [C]∥Proceedings of the Annual Conference of International Speech Communication Association. Baixs, France: ISCA, 2012: 14-17.

PAULIK M. Lattice-based training of bottleneck feature extraction neural networks [C]∥Proceedings of the Annual Conference of International Speech Communication Association. Baixs, France: ISCA, 2013: 89-93.

LIU D Y, WEI S, GUO W, et al. Lattice based optimization of bottleneck feature extractor with linear transformation [C]∥Proceedings of the International Conference on Audio, Speech and Signal Processing. Piscataway, NJ, USA: IEEE, 2014: 5617-5621.

YU D, SELTZER M L. Improved bottleneck features using pretrained deep neural networks [C]∥Proceedings of the Annual Conference of International Speech Communication Association. Baixs, France: ISCA, 2011: 237-240.

HOFFMEISTER B, KLEIN T, SCHLUTER R, et al. Frame based system combination and a comparison with weighted ROVER and CNC [C]∥Proceedings of the International Conference on Spoken Language Processing. Piscataway, NJ, USA: IEEE, 2006: 537-540.

ZOU H, HASTIE T. Regularization and variable selection via the elastic net [J]. Journal of the Royal Statistical Society: Series B Statistical Methodology, 2005, 67(2): 301-320.

BECK A, TEBOULLE M. A fast iterative shrinkage thresholding algorithm for linear inverse problems [J]. SIAM Journal on Imaging Sciences, 2009, 2(1): 183-202.[16] CHEN X, ZHAO Y. Building acoustic model ensembles by data sampling with enhanced trainings and features [J]. IEEE Transactions on Audio Speech and Language Processing, 2013, 21(3): 498-507.

POVEY D, KANEVSKY D, KINGSBURY B, et al. Boosted MMI for model and feature space discriminative training [C]∥Proceedings of the International Conference on Acoustics, Speech and Signal Processing. Piscataway, NJ, USA: IEEE, 2008: 4057-4060.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

A Dynamic Layer-Wise Gradient Sparsity and Gradient Merging Optimization Method for Deep Neural Networks

Learning New Words for Pronunciation Lexicon from Audio Data

I/O Performance Modeling and Prediction Method for Large-Scale Parallel Applications Using Machine Learning

Review of the Application of EBSILON Software in Combined Heat and Power System Optimization

Influence of Structural Parameters on the Summer Operational Characteristics of a Three-Tower-in-One Indirect Air-Cooling System

Related Author

JU Tao

KANG Heting

LIU Shuai

HUO Jiuyuan

范正光

屈丹

闫红刚

张文林

Related Institution

(School of Electronic and Information Engineering, Lanzhou Jiaotong University,,)

Institute of Information System Engineering, PLA Information Engineering University

⁰