Abstract

People with speech-language disorders often have limited vocabulary and/or inaccurate pronunciation. The potential benefit of speech-language technology for such people is manifold. For example, one may want to monitor over time the vocabulary of children with suspected developmental language disorders, or changes of pronunciation accuracy in response to speech-language therapy. In this topic, we plan to evaluate the performance of state-of-the-art approaches to pronunciation assessment with (i) remote recordings of English speaking people with speech pathologies, and (ii) bilingually raised German learning children with suspected developmental language disorders.

Your Tasks

• literature review
• making of (or support with) speech audio recordings (to be taken in Vienna)
• annotation of the recordings with respect to pronounciation accuracy
• training and testing of state-of-the-art approaches to automatic pronounciation assessment
• reporting your results (thesis writing, optional: paper)

Your Profile

• good knowledge of sound engineering, speech communication, phonetics
• good knowledge of Python
• interest in working with children is a plus

Contact

Philipp Aichinger (philipp.aichinger@meduniwien.ac.at) Barbara Schuppler (b.schuppler@tugraz.at)