Karjol, P and Ghosh, PK (2019) Broad phoneme class specific deep neural network based speech enhancement. In: 12th International Conference on Signal Processing and Communications, SPCOM 2018, 16 - 19 July 2018, Bangalore, pp. 372-376.
![]() |
PDF
SPCOM_2018.pdf - Published Version Restricted to Registered users only Download (299kB) | Request a copy |
Abstract
In this work, we present a speech enhancement algorithm using broad phoneme class specific deep neural networks (DNNs). We build a classifier network to obtain the probabilities of each class in every test frame. The enhanced speech is obtained as the linear combination of the outputs from these class specific DNNs with combination weights as the probabilities predicted by the classifier network. We experiment with two (vowel and non-vowel) and four (vowel, stop, fricative, nasal) broad phoneme classes. Experiments are performed using speech data from TIMIT corpus, nine noise types (four seen and five unseen), and four SNR conditions. Experimental results show that the proposed class specific DNN approach outperforms the single DNN based speech enhancement scheme for both seen and unseen noise types in all SNR conditions considered.
Item Type: | Conference Paper |
---|---|
Publication: | SPCOM 2018 - 12th International Conference on Signal Processing and Communications |
Publisher: | Institute of Electrical and Electronics Engineers Inc. |
Additional Information: | The copyright for this article belongs to Institute of Electrical and Electronics Engineers Inc. |
Keywords: | Linguistics; Signal processing; Signal to noise ratio; Speech enhancement, Combination weights; Linear combinations; Noise types; Phoneme class; Speech data; Speech enhancement algorithm, Deep neural networks |
Department/Centre: | Division of Electrical Sciences > Electrical Communication Engineering Division of Electrical Sciences > Electrical Engineering |
Date Deposited: | 08 Aug 2022 06:31 |
Last Modified: | 08 Aug 2022 06:31 |
URI: | https://eprints.iisc.ac.in/id/eprint/75487 |
Actions (login required)
![]() |
View Item |