Parkinson’s Disease Speech Analysis – Machine Learning based Evaluation of Speech Impairment (en)

* Presenting author
Day / Time: 09.03.2023, 10:40-11:00
Room: Saal Y8
Typ: Vortrag (strukturierte Sitzung)
Abstract: Parkinson’s disease is one of the most common neurodegenerative disorders. Speech impairments, caused by motor and non-motor deficits, are identified as frequently appearing and early symptoms. To track the disease’s progress and to evaluate the success of treatment, logopaedic assessments are performed by physicians. Results have to be tracked by themselves, mostly resulting in taking subjective notes.This paper presents a hierarchical approach to analyze the articulation, phonation, and prosody of vowels, composing the intelligibility of one’s speech. The three-step scheme consists of using speech samples of standalone vowels, vowels embedded in words, and text-bound speech, which are recorded and analyzed in an automated fashion. Acoustic features of each segment and differences in the signals and features of the different speech tasks are extracted and used to train a neural network. The trained model is used to provide a distinguished and objective assessment of the severity of a person’s speech intelligibility in real-time. The results as well as the recorded speech data is summarized in a report to assist the speech therapists work and enable tracking of speech intelligibility and progress of speech treatments over time.


strpos(): Passing null to parameter #1 ($haystack) of type string is deprecated

/kunden/431424_13355/webseiten/subdomains/, line 35
32.              */
33.             if (headers_sent()) {
34.                 $encoding = false;
35.             } elseif (strpos($HTTP_ACCEPT_ENCODING, 'x-gzip') !== false) {
36.                 $encoding = 'x-gzip';
37.             } elseif (strpos($HTTP_ACCEPT_ENCODING, 'gzip') !== false) {
38.                 $encoding = 'gzip';