Classroom lecture recognition

Authors:
Isabel Trancoso;Ricardo Nunes;Luís Neves
Affiliations:
INESC ID / IST, Lisbon, Portugal;INESC ID / IST, Lisbon, Portugal;INESC ID / IST, Lisbon, Portugal
Venue:
PROPOR'06 Proceedings of the 7th international conference on Computational Processing of the Portuguese Language
Year:
2006

Citing 4
Cited 1

Speech repairs, intonational phrases, and discourse markers: modeling speakers' utterances in spoken dialogue

Computational Linguistics
A TAG-based noisy channel model of speech repairs

ACL '04 Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics
Developments in continuous speech dictation using the 1995 ARPA NAB news task

ICASSP '96 Proceedings of the Acoustics, Speech, and Signal Processing, 1996. on Conference Proceedings., 1996 IEEE International Conference - Volume 01
Analysis and processing of lecture audio data: preliminary investigations

SpeechIR '04 Proceedings of the Workshop on Interdisciplinary Approaches to Speech Indexing and Retrieval at HLT-NAACL 2004

Simultaneous translation of lectures and speeches

Machine Translation

Quantified Score

Hi-index	0.00

Visualization

Abstract

The main goal of this work is to provide automatic transcriptions of classroom lectures for e-learning and e-inclusion applications. The first experiments using a recognition system trained for Broadcast News resulted in word error rates near 60%, clearly confirming the need for adaptation to the specific topic of the lectures, on one hand, and for better strategies for handling spontaneous speech. This paper describes the different domain adaptation steps that lowered the error rate to 45%, with very little transcribed adaptation material. It also includes a qualitative analysis of the different types of error, focusing on the ones related to a very high rate of disfluencies.