Comparing Human and Automatic Recognition of Dutch Dysarthric Continuous Speech: A Case Study
arXiv:2606.30237v1 Announce Type: new Abstract: In our goal to develop personalised dysarthric speech recognition (DSR) models, this study compared the recognition performances of human listeners and those of three state-of-the-art, off-the-shelf ASR systems (Whisper-large-V3, Google Chirp 3, and Omnilingual) on the recognition of Dutch continuous read and spontaneous speech from a single speaker with severe dysarthria. Results showed that both humans listeners and the three off-the-shelf ASR systems exhibit word error rates (WER) exceeding 70% on average, indicating that DSR is highly challen
This study, published in 2026, reflects ongoing research into the performance limitations of state-of-the-art ASR systems for highly challenging speech recognition tasks.
It highlights a significant frontier for AI improvement, specifically in accessibility and inclusivity for individuals with severe speech impairments, indicating current ASR technology is not universally robust.
The understanding that even advanced ASR systems require substantial improvement and personalization to effectively serve populations with severe dysarthric speech, opening avenues for dedicated research and development.
- · Researchers in personalized ASR
- · Assistive technology developers
- · Speech-language pathologists
- · Individuals with dysarthria
- · General-purpose ASR companies relying solely on off-the-shelf models for niche a
- · Users expecting seamless ASR for severe dysarthria
Major ASR companies will likely invest more in fine-tuning and specialized models for diverse speech patterns.
Development of personalized, on-device AI models for speech recognition will accelerate, enhancing accessibility for specific user groups.
Improved DSR could lead to greater social integration and employment opportunities for individuals with severe dysarthria, reshaping assistive communication technology markets.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL