Loading [MathJax]/extensions/MathZoom.js
Boosting of Prosodic and Pronunciation Features to Detect Mispronunciations of Non-Native Children | IEEE Conference Publication | IEEE Xplore

Boosting of Prosodic and Pronunciation Features to Detect Mispronunciations of Non-Native Children


Abstract:

Commercial products that support L2-learners with computer assisted pronunciation training usually focus per exercise only on one possible pronunciation mistake that is t...Show More

Abstract:

Commercial products that support L2-learners with computer assisted pronunciation training usually focus per exercise only on one possible pronunciation mistake that is typical for speakers of the respective L1 group. Acoustic models for words with wrong pronunciation are added to the system. In the present paper a more general approach with features that have proved to be widely independent of the learners' mother tongue is proposed. It is able to take various possible mistakes into consideration all at once. High dimensional feature vectors that encode prosodic varieties and differences of reference and recognized sentences are analyzed. With the ADABOOST algorithm those features are found, which contain the most important information to assess German children learning English. With 35 features 89 % of the agreement of experts is achieved.
Date of Conference: 15-20 April 2007
Date Added to IEEE Xplore: 04 June 2007
ISBN Information:

ISSN Information:

Conference Location: Honolulu, HI, USA

1. INTRODUCTION

A lot of research has been focused on computer assisted pronunciation training (CAPT) in the recent years which supports in most cases learners of the second language (L2) English. Primarily, language specific, rule based approaches have been investigated; all rules depend on the learners’ mother tongue. For the European languages, an important research project was ISLE (described by Herron et al.[1]), that has focused on adult German and Italian learners. In the European project Pf-Star [2] we focused on speech technologies for children. Based on these technologies, different systems for the assessment of pronunciation have been developed in our institute, e.g to objectively evaluate speech disordered children with a cleft lip and palate. Caller is a system for computer assisted language learning from Erlangen [3], a client/server system that can be started in a browser; speech is analyzed on a server, e.g. placed in a school's computer room (Fig. 1).

References

References is not available for this document.