Cross-Learning Fine-Tuning Strategy for Dysarthric Speech Recognition Via CDSD database
This addresses the problem of accurate speech recognition for individuals with dysarthria, offering a more efficient and generalizable approach, though it is incremental as it builds on existing fine-tuning methods.
The paper tackled dysarthric speech recognition by proposing a multi-speaker fine-tuning strategy, which improved individual speech pattern recognition and achieved up to 13.15% lower word error rate compared to single-speaker fine-tuning.
Dysarthric speech recognition faces challenges from severity variations and disparities relative to normal speech. Conventional approaches individually fine-tune ASR models pre-trained on normal speech per patient to prevent feature conflicts. Counter-intuitively, experiments reveal that multi-speaker fine-tuning (simultaneously on multiple dysarthric speakers) improves recognition of individual speech patterns. This strategy enhances generalization via broader pathological feature learning, mitigates speaker-specific overfitting, reduces per-patient data dependence, and improves target-speaker accuracy - achieving up to 13.15% lower WER versus single-speaker fine-tuning.