Using Phonological Context for Improved Recognition of Dysarthric Speech

Background and Application Domain Dysarthrias are a family of motor speech disorders which arise from neurological trauma, cerebro-vascular and conditions such as cerebral palsy. Impaired motor production in dysarthric speakers interferes with their basic process of speech production such as phonation, articulation and prosody. Hence their ability to communicate is severely restricted and drastically affects their quality of life. Even though speech therapy can help dysarthric speakers improve intelligibility; it cannot be expected to restore normal speech. Despite their unintelligible speech, dysarthric speakers prefer spoken interaction to other modalities for social communication. Hence some form of machine-assisted speech communication using a portable ASR system would be highly desirable.