Robust Speech Recognition Based on Discriminative Learning of Environmental Features

HAN Ji-qing; GAO Wen

您当前的位置：

首页 >

文章列表页 >

Robust Speech Recognition Based on Discriminative Learning of Environmental Features

更新时间：2025-07-16

- Robust Speech Recognition Based on Discriminative Learning of Environmental Features
- Acta Electronica Sinica Vol. 29, Issue 2, Pages: 196-198(2001)
- 作者机构：
  
  哈尔滨工业大学计算机科学与工程系,哈尔滨,150001
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP391.42
- Published：2001
- 稿件说明：
移动端阅览
HAN Ji-qing, GAO Wen. Robust Speech Recognition Based on Discriminative Learning of Environmental Features[J]. Acta Electronica Sinica, 2001, 29(2): 196-198.
DOI：

HAN Ji-qing, GAO Wen. Robust Speech Recognition Based on Discriminative Learning of Environmental Features[J]. Acta Electronica Sinica, 2001, 29(2): 196-198. DOI：

摘要

提出一种基于环境特征判别学习的顽健语音识别方法

它首先通过使用一个简单的分类器和梯度下降法迭代地学得环境特征

接着利用得到的环境特征从观测到的混噪语音特征中估计出纯净的语音特征

然后将估计出来的纯净语音特征用到后端的HMM分类器中.使用所提出的方法对不特定话者小词表进行实验

其系统误识率与基本HMM系统相比下降了33.3%.

Abstract

A discriminative learning method of environmental features is proposed for robust speech recognition

in which a simple classifier and the gradient descent algorithm are firstly used to iteratively learn the environmental features

and then the pure speech features are estimated from the observed speech features.Finally the estimations of the pure speech features are used in the back-end HMM classifier.Using the proposed method

a 33.3% reduction in word error rate is obtained relative to conventional HMM system for a speaker-independent

small vocabulary recognition task.

关键词

Keywords

references

Views

942

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research Progress of Stressed Speech Processing

Related Author

ZHANG Lei

HAN Ji-qing

WANG Cheng-fa

Related Institution

No data

⁰