Deep neural network-based linear predictive parameter estimations for speech enhancement

作者：Li Yaxing; Kang Sangwon^*

来源：IET Signal Processing, 2017, 11(4): 469-476.

DOI：10.1049/iet-spr.2016.0477

• speech enhancement
• parameter estimation
• neural nets
• maximum likelihood estimation
• Wiener filters
• regression analysis
• splines (mathematics)
• deep neural network
• linear predictive parameter estimation
• speech enhancement technique
• LP parameter estimation method
• DNN
• maximum-likelihood scheme
• time-smoothed Wiener filter
• performance evaluation
• log spectral distance evaluation
• composite multivariate adaptive regression spline modelling
• signal-to-noise ratio

This study presents a speech enhancement technique to improve noise corrupted speech via deep neural network (DNN)-based linear predictive (LP) parameter estimations of speech and noise. With regard to the LP coefficient estimation, an enhanced estimation method using a DNN with multiple layers was proposed. Excitation variances were then estimated via a maximum-likelihood scheme using observed noisy speech and estimated LP coefficients. A time-smoothed Wiener filter was further introduced to improve the enhanced speech quality. Performance was evaluated via log spectral distance, a composite multivariate adaptive regression splines modelling-based measure, and a segmental signal-to-noise ratio. The experimental results revealed that the proposed scheme outperformed competing methods.

• 出版日期2017-6