深度神經網路增強助聽器對中文聲調類別辨識的影響

// Effects of Deep Neural Network-Enhanced Hearing Aids on Categorical Perception of Mandarin Tones

本研究評估深度神經網絡（DNN）噪音消除對聽障者聽覺輔助器中普通話聲調辨識的影響｜使用DNN降噪的聽障者，在喧鬧環境下聲調辨識能力接近正常聽力者｜未使用DNN時，聽障者聲調識別明顯下降，顯示DNN技術有助於保留關鍵聲調資訊，提升噪音中的語音理解能力。

// This study evaluated how deep neural network (DNN) noise reduction in hearing aids affects Mandarin tone perception in hard of hearing (HH) listeners amid cafeteria noise. HH listeners using hearing aids with DNN noise reduction showed tone identification performance similar to normal-hearing listeners, especially at challenging noise levels. Without DNN processing, HH listeners had poorer tone discrimination in noise. The findings suggest that DNN noise reduction preserves critical tonal cues, improving speech perception for tonal language users in noisy environments.

Published 2026年7月29日