Automatic speech recognition in the diagnosis of primary progressive aphasia. Fraser, K. C., Rudzicz, F., Graham, N., & Rochon, E. In Proceedings of SLPAT 2013, 4th Workshop on Speech and Language Processing for Assistive Technologies, pages 47–54, Grenoble, France, 2013. abstract bibtex Narrative speech can provide a valuable source of information about an individual's linguistic abilities across lexical, syntactic, and pragmatic levels. However, analysis of narrative speech is typically done by hand, and is therefore extremely time-consuming. Use of automatic speech recognition (ASR) software could make this type of analysis more efficient and widely available. In this paper, we present the results of an initial attempt to use ASR technology to generate transcripts of spoken narratives from participants with semantic dementia (SD), progressive nonfluent aphasia (PNFA), and healthy controls. We extract text features from the transcripts and use these features, alone and in combination with acoustic features from the speech signals, to classify transcripts as patient versus control, and SD versus PNFA. Additionally, we generate artificially noisy transcripts by applying insertions, substitutions, and deletions to manually-transcribed data, allowing experiments to be conducted across a wider range of noise levels than are pro- duced by a tuned ASR system. We find that reasonably good classification accuracies can be achieved by selecting appropriate features from the noisy transcripts. We also find that the choice of using ASR data or manually transcribed data as the training set can have a strong effect on the accuracy of the classifiers.
@InProceedings{ fraser2013a,
author = {Kathleen C. Fraser and Frank Rudzicz and Naida Graham and
Elizabeth Rochon},
title = {Automatic speech recognition in the diagnosis of primary
progressive aphasia},
address = {Grenoble, France},
booktitle = {Proceedings of SLPAT 2013, 4th Workshop on Speech and
Language Processing for Assistive Technologies},
pages = {47--54},
year = {2013},
download = {http://ftp.cs.toronto.edu/pub/gh/Fraser-etal-SLPAT-2013.pdf}
,
abstract = {Narrative speech can provide a valuable source of
information about an individual's linguistic abilities
across lexical, syntactic, and pragmatic levels. However,
analysis of narrative speech is typically done by hand, and
is therefore extremely time-consuming. Use of automatic
speech recognition (ASR) software could make this type of
analysis more efficient and widely available. In this
paper, we present the results of an initial attempt to use
ASR technology to generate transcripts of spoken narratives
from participants with semantic dementia (SD), progressive
nonfluent aphasia (PNFA), and healthy controls. We extract
text features from the transcripts and use these features,
alone and in combination with acoustic features from the
speech signals, to classify transcripts as patient versus
control, and SD versus PNFA. Additionally, we generate
artificially noisy transcripts by applying insertions,
substitutions, and deletions to manually-transcribed data,
allowing experiments to be conducted across a wider range
of noise levels than are pro- duced by a tuned ASR system.
We find that reasonably good classification accuracies can
be achieved by selecting appropriate features from the
noisy transcripts. We also find that the choice of using
ASR data or manually transcribed data as the training set
can have a strong effect on the accuracy of the
classifiers.}
}
Downloads: 0
{"_id":{"_str":"53d57a48f414ae191e000395"},"__v":0,"authorIDs":[],"author_short":["Fraser, K. C.","Rudzicz, F.","Graham, N.","Rochon, E."],"bibbaseid":"fraser-rudzicz-graham-rochon-automaticspeechrecognitioninthediagnosisofprimaryprogressiveaphasia-2013","bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["Kathleen","C."],"propositions":[],"lastnames":["Fraser"],"suffixes":[]},{"firstnames":["Frank"],"propositions":[],"lastnames":["Rudzicz"],"suffixes":[]},{"firstnames":["Naida"],"propositions":[],"lastnames":["Graham"],"suffixes":[]},{"firstnames":["Elizabeth"],"propositions":[],"lastnames":["Rochon"],"suffixes":[]}],"title":"Automatic speech recognition in the diagnosis of primary progressive aphasia","address":"Grenoble, France","booktitle":"Proceedings of SLPAT 2013, 4th Workshop on Speech and Language Processing for Assistive Technologies","pages":"47–54","year":"2013","download":"http://ftp.cs.toronto.edu/pub/gh/Fraser-etal-SLPAT-2013.pdf","abstract":"Narrative speech can provide a valuable source of information about an individual's linguistic abilities across lexical, syntactic, and pragmatic levels. However, analysis of narrative speech is typically done by hand, and is therefore extremely time-consuming. Use of automatic speech recognition (ASR) software could make this type of analysis more efficient and widely available. In this paper, we present the results of an initial attempt to use ASR technology to generate transcripts of spoken narratives from participants with semantic dementia (SD), progressive nonfluent aphasia (PNFA), and healthy controls. We extract text features from the transcripts and use these features, alone and in combination with acoustic features from the speech signals, to classify transcripts as patient versus control, and SD versus PNFA. Additionally, we generate artificially noisy transcripts by applying insertions, substitutions, and deletions to manually-transcribed data, allowing experiments to be conducted across a wider range of noise levels than are pro- duced by a tuned ASR system. We find that reasonably good classification accuracies can be achieved by selecting appropriate features from the noisy transcripts. We also find that the choice of using ASR data or manually transcribed data as the training set can have a strong effect on the accuracy of the classifiers.","bibtex":"@InProceedings{\t fraser2013a,\n author\t= {Kathleen C. Fraser and Frank Rudzicz and Naida Graham and\n\t\t Elizabeth Rochon},\n title\t\t= {Automatic speech recognition in the diagnosis of primary\n\t\t progressive aphasia},\n address\t= {Grenoble, France},\n booktitle\t= {Proceedings of SLPAT 2013, 4th Workshop on Speech and\n\t\t Language Processing for Assistive Technologies},\n pages\t\t= {47--54},\n year\t\t= {2013},\n download\t= {http://ftp.cs.toronto.edu/pub/gh/Fraser-etal-SLPAT-2013.pdf}\n\t\t ,\n abstract\t= {Narrative speech can provide a valuable source of\n\t\t information about an individual's linguistic abilities\n\t\t across lexical, syntactic, and pragmatic levels. However,\n\t\t analysis of narrative speech is typically done by hand, and\n\t\t is therefore extremely time-consuming. Use of automatic\n\t\t speech recognition (ASR) software could make this type of\n\t\t analysis more efficient and widely available. In this\n\t\t paper, we present the results of an initial attempt to use\n\t\t ASR technology to generate transcripts of spoken narratives\n\t\t from participants with semantic dementia (SD), progressive\n\t\t nonfluent aphasia (PNFA), and healthy controls. We extract\n\t\t text features from the transcripts and use these features,\n\t\t alone and in combination with acoustic features from the\n\t\t speech signals, to classify transcripts as patient versus\n\t\t control, and SD versus PNFA. Additionally, we generate\n\t\t artificially noisy transcripts by applying insertions,\n\t\t substitutions, and deletions to manually-transcribed data,\n\t\t allowing experiments to be conducted across a wider range\n\t\t of noise levels than are pro- duced by a tuned ASR system.\n\t\t We find that reasonably good classification accuracies can\n\t\t be achieved by selecting appropriate features from the\n\t\t noisy transcripts. We also find that the choice of using\n\t\t ASR data or manually transcribed data as the training set\n\t\t can have a strong effect on the accuracy of the\n\t\t classifiers.}\n}\n\n","author_short":["Fraser, K. C.","Rudzicz, F.","Graham, N.","Rochon, E."],"key":"fraser2013a","id":"fraser2013a","bibbaseid":"fraser-rudzicz-graham-rochon-automaticspeechrecognitioninthediagnosisofprimaryprogressiveaphasia-2013","role":"author","urls":{},"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"www.cs.toronto.edu/~fritz/tmp/compling.bib","creationDate":"2014-07-27T22:16:40.323Z","downloads":0,"keywords":[],"search_terms":["automatic","speech","recognition","diagnosis","primary","progressive","aphasia","fraser","rudzicz","graham","rochon"],"title":"Automatic speech recognition in the diagnosis of primary progressive aphasia","year":2013,"dataSources":["n8jB5BJxaeSmH6mtR","6b6A9kbkw4CsEGnRX"]}