Szczegóły publikacji
Opis bibliograficzny
Monitoring of Parkinson’s disease progression based on speech signal / Daria HEMMERLING, Magdalena Wójcik-Pędziwiatr, Paweł Jaciów, Bartosz Ziółko, Magdalena IGRAS-CYBULSKA // W: ICICT 2023 [Dokument elektroniczny] : 2023 6th International Conference on Information and Computer Technologies : Raleigh, USA, 24–26 March 2023 : proceedings. — Wersja do Windows. — Dane tekstowe. — Piscataway : IEEE, cop. 2023. — ( International Conference on Information and Computer Technologies ; ISSN 2769-4542 ). — Dod. ISBN: 979-8-3503-0096-3. — e-ISBN: 979-8-3503-0095-6. — S. 132–137. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. 137, Abstr. — Publikacja dostępna online od: 2023-08-30. — B. Ziółko - brak afiliacji AGH
Autorzy (5)
- AGHHemmerling Daria
- Wójcik-Pędziwiatr Magdalena
- Jaciów Paweł
- Ziółko Bartosz
- AGHIgras-Cybulska Magdalena
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 148435 |
|---|---|
| Data dodania do BaDAP | 2023-09-22 |
| Tekst źródłowy | URL |
| DOI | 10.1109/ICICT58900.2023.00029 |
| Rok publikacji | 2023 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Institute of Electrical and Electronics Engineers (IEEE) |
| Czasopismo/seria | International Conference on Information and Computer Technologies |
Abstract
The goal of this study was to determine whether the speech analysis might be useful tool to track the motor symptoms of patients with Parkinson’s disease during the levodopa treatment with reference to The Unified Parkinson’s Disease Rating Scale (UPDRS). Due to speech impairments caused PD we carried out the acoustic analysis based on automatic speech recognition system to extract vowels and consonant from speech signals to indicate the UPDRS score in an automatic manner. Speech analysis was the input to Support Vector Machine, Support Vector Regression and Gaussian Process Regression algorithms to determine UPDRS score. The lowest mean absolute error of UPDRS score achieved in this study is 8.52 points (ρ=0.72,r=0.69) when the analysis was carried out using only vowels extracted from speech signal. The developed method may help to follow the patient’s speech fluctuations and adjust the treatment.
