Automatic testing of speech recognition.
Speech reception tests are commonly administered by manually scoring the oral response of the subject. This requires a test supervisor to be continuously present. To avoid this, a subject can type the response, after which it can be scored automatically. However, spelling errors may then be counted...
| Publicado en: | International Journal of Audiology Vol. 48; no. 2; pp. 80 - 91 |
|---|---|
| Autores principales: | , , |
| Formato: | algorithm research tables/charts Journal Article |
| Publicado: |
Taylor & Francis Ltd
Feb2009
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=105458274&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 105458274 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 14992027 JW2 jtl: International Journal of Audiology issn: 14992027 maglogo: Y pubinfo: dt: Feb2009 vid: 48 iid: 2 pid: 377 pub: Taylor & Francis Ltd place: Philadelphia, Pennsylvania artinfo: ui: 105458274 2010200285 10.1080/14992020802400662 NLM19219692 105458274 ppf: 80 ppct: 11 formats: fmt: @attributes: type: P tig: atl: Automatic testing of speech recognition. aug: au: Francart T Moonen M Wouters J affil: ExpORL, Department of Neurosciences, Katholieke Universiteit Leuven, Leuven, Belgium. tom.francart@med.kuleuven.be sug: subj: Algorithms Evaluation Automation Speech Discrimination Tests Clinical Assessment Tools Descriptive Statistics Evaluation Research Funding Source Netherlands Spelling Human ab: Speech reception tests are commonly administered by manually scoring the oral response of the subject. This requires a test supervisor to be continuously present. To avoid this, a subject can type the response, after which it can be scored automatically. However, spelling errors may then be counted as recognition errors, influencing the test results. We demonstrate an autocorrection approach based on two scoring algorithms to cope with spelling errors. The first algorithm deals with sentences and is based on word scores. The second algorithm deals with single words and is based on phoneme scores. Both algorithms were evaluated with a corpus of typed answers based on three different Dutch speech materials. The percentage of differences between automatic and manual scoring was determined, in addition to the mean difference in speech recognition threshold. The sentence correction algorithm performed at a higher accuracy than commonly obtained with these speech materials. The word correction algorithm performed better than the human operator. Both algorithms can be used in practice and allow speech reception tests with open set speech materials over the internet. pubtype: Academic Journal doctype: algorithm research tables/charts Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|