2025
Prosody-aware dual-stream neural model for speaker-adaptive voice command recognition
A Prosody-Aware Dual-Stream Neural Model designed for speaker-adaptive voice command recognition, enabling robust classification of short voice commands in real-world conditions and confirming that the combination of acoustic and prosodic information provides a more stable and adaptive representation of speech.
Oleksii Matsiievskyi, I. Bosenko, Igor Achkasov et al.
· AI@DTESI · 0 citations