This paper investigates different trade-offs between the number of model parameters and enhanced speech qualities by employing several deep tensor-to-vector regression models for speech enhancement. We find that a hybrid architecture, namely CNN-TT, is capable of maintaining a good quality performance with a reduced model parameter size. CNN-TT is composed of several convolutional layers at the bottom for feature extraction to improve speech quality and a tensor-train (TT) output layer on the top to reduce model parameters. We first derive a new upper bound on the generalization power of the convolutional neural network (CNN) based vector-to-vector regression models. Then, we provide experimental evidence on the Edinburgh noisy speech corpus to demonstrate that, in single-channel speech enhancement, CNN outperforms DNN at the expense of a small increment of model sizes. Besides, CNN-TT slightly outperforms the CNN counterpart by utilizing only 32% of the CNN model parameters. Besides, further performance improvement can be attained if the number of CNN-TT parameters is increased to 44% of the CNN model size. Finally, our experiments of multi-channel speech enhancement on a simulated noisy WSJ0 corpus demonstrate that our proposed hybrid CNN-TT architecture achieves better results than both DNN and CNN models in terms of better-enhanced speech qualities and smaller parameter sizes.

Qi, J., Hu, H.u., Wang, Y., Yang, C.H., Siniscalchi, S.M., Lee, C. (2020). Exploring Deep Hybrid Tensor-to-Vector Network Architectures for Regression Based Speech Enhancement. In INTERSPEECH 2020 (pp. 76-80) [10.21437/Interspeech.2020-1900].

Exploring Deep Hybrid Tensor-to-Vector Network Architectures for Regression Based Speech Enhancement

Siniscalchi, Sabato Marco
Co-ultimo
Supervision
;
2020-01-01

Abstract

This paper investigates different trade-offs between the number of model parameters and enhanced speech qualities by employing several deep tensor-to-vector regression models for speech enhancement. We find that a hybrid architecture, namely CNN-TT, is capable of maintaining a good quality performance with a reduced model parameter size. CNN-TT is composed of several convolutional layers at the bottom for feature extraction to improve speech quality and a tensor-train (TT) output layer on the top to reduce model parameters. We first derive a new upper bound on the generalization power of the convolutional neural network (CNN) based vector-to-vector regression models. Then, we provide experimental evidence on the Edinburgh noisy speech corpus to demonstrate that, in single-channel speech enhancement, CNN outperforms DNN at the expense of a small increment of model sizes. Besides, CNN-TT slightly outperforms the CNN counterpart by utilizing only 32% of the CNN model parameters. Besides, further performance improvement can be attained if the number of CNN-TT parameters is increased to 44% of the CNN model size. Finally, our experiments of multi-channel speech enhancement on a simulated noisy WSJ0 corpus demonstrate that our proposed hybrid CNN-TT architecture achieves better results than both DNN and CNN models in terms of better-enhanced speech qualities and smaller parameter sizes.
2020
Settore ING-INF/05 - Sistemi Di Elaborazione Delle Informazioni
Qi, J., Hu, H.u., Wang, Y., Yang, C.H., Siniscalchi, S.M., Lee, C. (2020). Exploring Deep Hybrid Tensor-to-Vector Network Architectures for Regression Based Speech Enhancement. In INTERSPEECH 2020 (pp. 76-80) [10.21437/Interspeech.2020-1900].
File in questo prodotto:
File Dimensione Formato  
1900.pdf

accesso aperto

Tipologia: Versione Editoriale
Dimensione 969.64 kB
Formato Adobe PDF
969.64 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10447/636628
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 8
  • ???jsp.display-item.citation.isi??? 7
social impact