DMCA
The Cambridge University Spoken Document Retrieval System (1999)
| Citations: | 39 - 2 self |
BibTeX
@MISC{Johnson99thecambridge,
author = {Sue Johnson and Pierre Jourlin and Gareth Moore and Karen Spärck Jones and Philip Woodland},
title = {The Cambridge University Spoken Document Retrieval System},
year = {1999}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
This paper describes the spoken document retrieval system that we have been developing and assesses its performance using automatic transcriptions of about 50 hours of broadcast news data. The recognition engine is based on the HTK broadcast news transcription system and the retrieval engine is based on the techniques developed at City University. The retrieval performance over a wide range of speech transcription error rates is presented and a number of recognition error metrics that more accurately reflect the impact of transcription errors on retrieval accuracy are defined and computed. The results demonstrate the importance of high accuracy automatic transcription. The final system is currently being evaluated on the 1998 TREC-7 spoken document retrieval task.
Keyphrases
transcription error trec-7 spoken document retrieval task htk broadcast news transcription system automatic transcription broadcast news data high accuracy automatic transcription retrieval performance wide range speech transcription error rate retrieval accuracy city university retrieval engine final system recognition engine spoken document retrieval system recognition error metric
