A therapist-annotated dysarthric speech dataset and a three-stage ASR framework show that Whisper-large localizes mispronunciations precisely, with substitution errors detected best and 70.1% of ASR error descriptions matching therapist labels.
In this setup, both the speech recording and the corresponding target text are available, enabling precise evaluation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Towards Temporally Explainable Dysarthric Speech Clarity Assessment
A therapist-annotated dysarthric speech dataset and a three-stage ASR framework show that Whisper-large localizes mispronunciations precisely, with substitution errors detected best and 70.1% of ASR error descriptions matching therapist labels.