Speaker adaptation for recognition systems with a large vocabulary

Proceedings. Electrotechnical Conference Integrating Research, Industry and Education in Energy and Communication Engineering', Pub Date : 1989-04-11 DOI:10.1109/MELCON.1989.50027

F. Class, P. Regel, K. Trottler

{"title":"Speaker adaptation for recognition systems with a large vocabulary","authors":"F. Class, P. Regel, K. Trottler","doi":"10.1109/MELCON.1989.50027","DOIUrl":null,"url":null,"abstract":"Algorithms for a fast speaker adaptation in a speech-recognition system are described. The techniques aim at transformations of the feature vectors, which have to be optimized with respect to some constraints. The methods transform every feature vector, computed in a 10-ms frame rate, into a speaker-normalized vector. The advantage of adaptation by transforming the feature vectors is that this procedure can be applied no matter which classification scheme is used. It is shown that, by means of adaptation procedures based on statistical correlation analysis, error rates as low as those of a speaker-dependent recognition system can be achieved after an extremely short training phase with any new speaker. The key is that the feature vectors are extended nonlinearly to a polynomial vector of second or higher order. Since the algorithms necessary for calculating the transformation matrices are typical for signal processing a real-time implementation on digital signal processors appears feasible.<<ETX>>","PeriodicalId":380214,"journal":{"name":"Proceedings. Electrotechnical Conference Integrating Research, Industry and Education in Energy and Communication Engineering',","volume":"27 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1989-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"4","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings. Electrotechnical Conference Integrating Research, Industry and Education in Energy and Communication Engineering',","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/MELCON.1989.50027","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 4

Abstract

Algorithms for a fast speaker adaptation in a speech-recognition system are described. The techniques aim at transformations of the feature vectors, which have to be optimized with respect to some constraints. The methods transform every feature vector, computed in a 10-ms frame rate, into a speaker-normalized vector. The advantage of adaptation by transforming the feature vectors is that this procedure can be applied no matter which classification scheme is used. It is shown that, by means of adaptation procedures based on statistical correlation analysis, error rates as low as those of a speaker-dependent recognition system can be achieved after an extremely short training phase with any new speaker. The key is that the feature vectors are extended nonlinearly to a polynomial vector of second or higher order. Since the algorithms necessary for calculating the transformation matrices are typical for signal processing a real-time implementation on digital signal processors appears feasible.<>

查看原文本刊更多论文

词汇量大的识别系统中的说话人自适应

介绍了语音识别系统中快速自适应说话人的算法。这些技术的目标是特征向量的变换，必须根据一些约束条件对其进行优化。该方法将以10ms帧速率计算的每个特征向量转换为说话人归一化向量。通过变换特征向量进行自适应的优点是，无论采用哪种分类方案，该方法都可以适用。研究表明，通过基于统计相关分析的自适应过程，在对任何新说话人进行极短的训练阶段后，都可以达到与依赖说话人识别系统一样低的错误率。关键是将特征向量非线性地扩展为二阶或更高阶的多项式向量。由于计算变换矩阵所需的算法是信号处理的典型算法，因此在数字信号处理器上实时实现似乎是可行的。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Proceedings. Electrotechnical Conference Integrating Research, Industry and Education in Energy and Communication Engineering',

自引率

0.00%

发文量