{"title":"Investigation of Inspection Methods in Acoustic Analysis Using Pronunciation Feature Extraction","authors":"N. Yagi, Yutaka Hata, Y. Sakai","doi":"10.1109/ICMLC56445.2022.9941300","DOIUrl":null,"url":null,"abstract":"Since the structure of speech is wide-ranging such as prosody, articulation, vocalization, and breathing, there is no screening test for speech disorder unlike the areas of aphasia and dysphagia. Speech intelligibility in speech-language pathology is evaluated by a Speech-Language-Hearing Therapist (ST), however the evaluation time per person is long and the evaluation criteria are ambiguous. So, the evaluation results will differ depending on the ST. Therefore, in this study, we proposed a system to easily inspect the normality of pronunciation by using 8 characteristics of data divided into single notes. As the results, this system enabled to identify whether the pronunciation is normal or abnormal with high accuracy of 93.3 %.","PeriodicalId":117829,"journal":{"name":"2022 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 International Conference on Machine Learning and Cybernetics (ICMLC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICMLC56445.2022.9941300","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Since the structure of speech is wide-ranging such as prosody, articulation, vocalization, and breathing, there is no screening test for speech disorder unlike the areas of aphasia and dysphagia. Speech intelligibility in speech-language pathology is evaluated by a Speech-Language-Hearing Therapist (ST), however the evaluation time per person is long and the evaluation criteria are ambiguous. So, the evaluation results will differ depending on the ST. Therefore, in this study, we proposed a system to easily inspect the normality of pronunciation by using 8 characteristics of data divided into single notes. As the results, this system enabled to identify whether the pronunciation is normal or abnormal with high accuracy of 93.3 %.