Automatic Evaluation of Pathologic Speech - from Research to Routine Clinical Use

  • Previously we have shown that ASR technology can be used to objectively evaluate pathologic speech. Here we report on progress for routine clinical use: 1) We introduce an easy-to-use recording and evaluation environment. 2) We confirm our previous results for a larger group of patients. 3) We show that telephone speech can be analyzed with the same methods with only a small loss of agreement with human experts. 4) We show that prosodic information leads to more robust results. 5) We show that text reference instead of transliteration can be used for evaluation. Using word accuracy of a speech recognizer and prosodic features as features for SVM regression, we achieve a correlation of .90 between the automatic analysis and human experts.

Export metadata

Additional Services

Search Google Scholar
Metadaten
Author:Elmar Nöth, Andreas Maier, Tino Haderlein, Korbinian Riedhammer, Frank Rosanowski, Maria Schuster
Parent Title (English):TSD 2007, 10th International Conference on Text, Speech and Dialogue, Pilsen, Czech Republic, September 2007.
Publisher:Springer-Verlag
Document Type:Conference Proceeding
Language:English
Publication Year:2007
Tag:Automatic Evaluation; Automatic Speech Recognition; Pathologic Speech; Prosodic Feature; Speech Recognizer
First Page:294
Last Page:301
faculties / departments:Fakultät für Informatik
Dewey Decimal Classification:0 Informatik, Informationswissenschaft, allgemeine Werke / 00 Informatik, Wissen, Systeme / 000 Informatik, Informationswissenschaft, allgemeine Werke