Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | 장준혁 | - |
dc.date.accessioned | 2021-11-30T01:18:09Z | - |
dc.date.available | 2021-11-30T01:18:09Z | - |
dc.date.issued | 2020-05 | - |
dc.identifier.citation | DIGITAL SIGNAL PROCESSING, v. 102, article no. 102760 | en_US |
dc.identifier.issn | 1051-2004 | - |
dc.identifier.issn | 1095-4333 | - |
dc.identifier.uri | https://www.sciencedirect.com/science/article/pii/S1051200420301056?via%3Dihub | - |
dc.identifier.uri | https://repository.hanyang.ac.kr/handle/20.500.11754/166516 | - |
dc.description.abstract | In this paper, we propose a deep neural network (DNN) ensemble for reducing artificial noise in speech bandwidth extension (BWE). The proposed DNN ensemble consists of three DNN models; one is a classification model, and the other two are regression models. When estimating sub-band energies of the high-frequency region using sequential DNNs in a frequency domain, the over-estimation of sub-band energies causes annoying artificial noise. To mitigate this artificial noise, we design a DNN classification model that can classify over-estimation frames against normal frames. Then, we separately develop two DNN regression models using half of the entire training set and a limited training set built with overestimation frames and some normal frames to improve the performance at the over-estimation frames. Since the outputs of the classification model are probabilities of either a normal frame or an overestimation frame, respectively, two regression models are adjustably combined by using the probabilistic weights; thus, the final output of the DNN ensemble is the weighted sum of two estimated sub-band energies. As a result, artificial noise is significantly reduced, yielding improved speech quality. The proposed method is objectively and subjectively evaluated by comparing it with conventional approaches. | en_US |
dc.description.sponsorship | This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2017R1A2A1A17069651). | en_US |
dc.language.iso | en | en_US |
dc.publisher | ACADEMIC PRESS INC ELSEVIER SCIENCE | en_US |
dc.subject | Bandwidth extension | en_US |
dc.subject | Deep neural network | en_US |
dc.subject | Ensemble | en_US |
dc.subject | Artificial noise | en_US |
dc.title | Deep neural network ensemble for reducing artificial noise in bandwidth extension | en_US |
dc.type | Article | en_US |
dc.relation.volume | 102 | - |
dc.identifier.doi | 10.1016/j.dsp.2020.102760 | - |
dc.relation.page | 1-6 | - |
dc.relation.journal | DIGITAL SIGNAL PROCESSING | - |
dc.contributor.googleauthor | Noh, Kyoungjin | - |
dc.contributor.googleauthor | Chang, Joon-Hyuk | - |
dc.relation.code | 2020049761 | - |
dc.sector.campus | S | - |
dc.sector.daehak | COLLEGE OF ENGINEERING[S] | - |
dc.sector.department | SCHOOL OF ELECTRONIC ENGINEERING | - |
dc.identifier.pid | jchang | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.