eduzhai > Applied Sciences > Engineering >

Tdcgan Temporal Dilated Convolutional Generative Adversarial Network for End-to-end Speech Enhancement

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 5 pages

Abstract: In this paper, in order to further deal with the performance degradationcaused by ignoring the phase information in conventional speech enhancementsystems, we proposed a temporal dilated convolutional generative adversarialnetwork (TDCGAN) in the end-to-end based speech enhancement architecture. Forthe first time, we introduced the temporal dilated convolutional network withdepthwise separable convolutions into the GAN structure so that the receptivefield can be greatly increased without increasing the number of parameters. Wealso first explored the effect of signal-to-noise ratio (SNR) penalty item asregularization of the loss function of generator on improving the SNR ofenhanced speech. The experimental results demonstrated that our proposed methodoutperformed the state-of-the-art end-to-end GAN-based speech enhancement.Moreover, compared with previous GAN-based methods, the proposed TDCGAN couldgreatly decreased the number of parameters. As expected, the work alsodemonstrated that the SNR penalty item as regularization was more effectivethan $L1$ on improving the SNR of enhanced speech.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×