Speech analysis software as a clinical screening tool for speech sound disorders in Cantonese-speaking children
Refereed conference paper presented and published in conference proceedings


摘要Background: communicate within normal limits and those who may have a communicative disorder (Shipley & McAfee, 2008). People in the high-risk group could be seen or referred for comprehensive evaluation in an earlier time after being screened out.
In order to make speech screening more accessible to parents and teachers, the development of a Cantonese speech sound analyzer can be one of the ways. Speech sound analyzer is a computer program for acoustic analysis of speech sounds input into the program according to the features of the sound spectrum. In recent years, computerized digital sound analyzers are widely used in detecting voice disorders, speech and audio-signal processing, etc. There is no speech sound analysis program available for Cantonese speech. And this makes the development of application of computerized program in speech therapy difficult. Thus, the purpose of this study is to compare the accuracy of the screening results obtained by qualified speech therapists and by the sound-analysing system built by a database of correctly produced cantonese phonemes. It can also be investigated to see if the accuracy of the sound anaylzing system is high enough for screening children with speech sound disorders.
Methodology: 100 Cantonese-speaking children with aged 3-6 will be recruited from local kindergartens. Additional criteria includes case histories that were unremarkable in terms of birth, development and general health; no significant hearing problems, etc. A powerpoint with pictures of single words covering all Cantonese speech sounds including all consonants, vowels, diphthongs and tones will be used as the screening tool. The total length of screening is expected to be within 10 minutes. The whole process will be video-taped by a camera and audio-taped with a data-storage computer. The data collected will be used to build up the database of the speech sound analyzer in computer laboratory. All the samples will also be evaluated by qualified speech therapists by perceptual analysis, with at least 5 years of practice in the field. The results obtained from the speech therapists will be used as the gold standard to build up the database of correct sound productions. The results of the built speech sound analazer in the final stage will also be compared with that obtained from the speech therapists.
Results: implies that speech sound analysing technology is comparable with speech therapists in a screening of speech sound disorders. Computerized speech sound analyzer could be used as the screening tool to screen out children with speech sound disorders.
If the differences are significant, the unmatched pairs could be catagorized and analyzed accordingly to the sound patterns for further investigation on the limitation of speech sound analysing technology on Cantonese speech screening.
Conclusion: The research results will enable us to compare the accuracy of a speech screening test performed by a speech sound analyzer with that obtained from speech therapists systematically. This can also provide us more information about the feasibility and accuracy of the future application of sound-analyzing technology in speech therapy.
著者Ng C. W. Y., Lee K. Y, S., Lee T., Tong M. C. F.
會議名稱Hong Kong Speech and Hearing Symposium 2016
會議地點Hong Kong
會議論文集題名Hong Kong Speech and Hearing Symposium 2016
出版地Hong Kong
頁次102 - 103
關鍵詞speech therapy, Speech analysis software

上次更新時間 2018-22-01 於 05:59