Paper Info
Title
Towards a generic approach for automatic speech recognition error detection and classification
Abstract
Automatic Speech Recognition (ASR) errors are essentially unavoidable. This premise motivates the attempts to develop post hoc tools that tackle the ASR errors. This paper addresses the problem of errors in continuous speech recognition outputs to improve the exploitation of ASR transcriptions. We propose a generic classifier-based approach for both error detection and error type classification. Unlike the majority of research in this field, we handle the recognition errors independently from the ASR decoder using a set of features derived exclusively from the recognizer output and hence should be usable with any ASR system. As a result, experiments on TV program transcription data have shown that the proposed non-decoder features setup leads to achieve competitive performances, compared to state of the art systems, in ASR error detection and classification. Furthermore, we have shown that Support Vector Machines trained on the proposed features set appear to be an effective classifier for the ASR error type classification with an Accuracy of 82.41%.
Year
DOI
Venue
2018
10.1109/ATSIP.2018.8364511
2018 4th International Conference on Advanced Technologies for Signal and Image Processing (ATSIP)
Keywords
Field
DocType
generic approach,automatic speech recognition error detection,post hoc tools,ASR errors,continuous speech recognition outputs,ASR transcriptions,ASR decoder,ASR system,ASR error detection,ASR error type classification,support vector machines
USable,Transcription (linguistics),Computer science,Support vector machine,Error detection and correction,Speech recognition,Classifier (linguistics)
Conference
ISBN
Citations 
PageRank 
978-1-5386-5240-4
0
0.34
References 
Authors
13
4
Name
Order
Citations
PageRank
Rahhal Errattahi121.08
El Hannani24811.48
Thomas Hain348639.70
Hassan Ouahmane412.73