语音增强基于小脑模型控制器-associationforcomputational.pdfVIP

  • 4
  • 0
  • 约3.48万字
  • 约 13页
  • 2017-10-06 发布于天津
  • 举报

语音增强基于小脑模型控制器-associationforcomputational.pdf

语音增强基于小脑模型控制器-associationforcomputational

The 2015 Conference on Computational Linguistics and Speech Processing ROCLING 2015, pp. 123-135  The Association for Computational Linguistics and Chinese Language Processing 語音增強基於小腦模型控制器 朱皓駿 Hao-Chun Chu ,李仲溪 Jung-Hsi Lee 方士豪 Shih-Hau Fang ,林志民 Chih-Min Lin 元智大學電機工程學系 Department of Electrical Engineering, Yuan Ze University david4633221@ {eejlee, shfang, cml}@.tw 張雲帆 Yun-Fan Chang , 曹昱 Yu Tsao 中央研究院資訊科技創新研究中心 Research Center for Information Technology Innovation, Academia Sinica {she2113, yu.tsao}@.tw 摘要 本文提出了一個小腦模型控制器(Cerebellar Model Articulation Controller, CMAC)應用於 語音增強系統(Speech Enhancement System) ,所提出的CMAC 使用歸一化梯度下降法 (Normalized Gradient Descent Method) 增加CMAC 參數的自適應學習速度,具有比傳統 類神經網路方法更快的學習速度、體積小且良好的泛化,因此更適合做高速的訊號處理。 實驗方面,使用CMAC 與MMSE 做比較,為了比較性能 ,我們用了三種語音評估方法 來做CMAC 消除雜音及MMSE 消除雜音後的數值比較,分別為(Perceptual Evaluation of Speech Quality, PESQ) 、(Segmental Signal-to-Noise Ratio, SSNR)以及(Speech Distortion Index, SDI) 。由實驗結果可知,在三種評估方法,CMAC 皆能達到較佳的結果。 Abstract Traditionally, cerebellar model articulation controller (CMAC) is used in motor control, inverted pendulum robot, and nonlinear channel equalization. In this study, we investigate the capability of CMAC for speech enhancement. We construct a CMAC-based supervised speech enhancement system, which includes offline and online phases. In the offline phase, a paired noisy-clean speech dataset is prepared and used to train the parameters in a CMAC model. In the online phase, the trained CMAC model transforms the input noisy speech

文档评论(0)

1亿VIP精品文档

相关文档