Результаты исследований: Публикации в книгах, отчётах, сборниках, трудах конференций › статья в сборнике материалов конференции › научная › Рецензирование
Using Convolutional Neural Networks to Restore Audiosignals. / Morozova, Kristina; Rakitskiy, Anton.
2021 IEEE 22nd International Conference of Young Professionals in Electron Devices and Materials, EDM 2021 - Proceedings. IEEE Computer Society, 2021. стр. 524-527 9507633 (International Conference of Young Specialists on Micro/Nanotechnologies and Electron Devices, EDM; Том 2021-June).Результаты исследований: Публикации в книгах, отчётах, сборниках, трудах конференций › статья в сборнике материалов конференции › научная › Рецензирование
}
TY - GEN
T1 - Using Convolutional Neural Networks to Restore Audiosignals
AU - Morozova, Kristina
AU - Rakitskiy, Anton
N1 - Publisher Copyright: © 2021 IEEE.
PY - 2021/6/30
Y1 - 2021/6/30
N2 - In this article, we will explore the possibility of using neural networks to solve the problem of audio signal recovery. Based on the previously obtained results of using convolutional neural networks for the extraction of the voice part, the concept of a convolutional neural network was developed, designed to correct a distorted audio signal. This article presents the initial concept of this neural network architecture, which unfortunately gave unsatisfactory results. Nevertheless, based on the concept of this network, several new neural network architectures have been developed, specifically focused on the restoration of distorted audio signal, but at the same time, the shortcomings of the basic architecture were taken into account. The article contains descriptions of all these architectures and the results of their application to restore the drummer part in a musical composition from which it was deleted. We also studied the effect of increasing the number of neural networks on the efficiency of signal recovery.
AB - In this article, we will explore the possibility of using neural networks to solve the problem of audio signal recovery. Based on the previously obtained results of using convolutional neural networks for the extraction of the voice part, the concept of a convolutional neural network was developed, designed to correct a distorted audio signal. This article presents the initial concept of this neural network architecture, which unfortunately gave unsatisfactory results. Nevertheless, based on the concept of this network, several new neural network architectures have been developed, specifically focused on the restoration of distorted audio signal, but at the same time, the shortcomings of the basic architecture were taken into account. The article contains descriptions of all these architectures and the results of their application to restore the drummer part in a musical composition from which it was deleted. We also studied the effect of increasing the number of neural networks on the efficiency of signal recovery.
KW - audio signal
KW - machine learning methods
KW - neural networks
KW - regression
KW - signal recovery
UR - http://www.scopus.com/inward/record.url?scp=85113531351&partnerID=8YFLogxK
U2 - 10.1109/EDM52169.2021.9507633
DO - 10.1109/EDM52169.2021.9507633
M3 - Conference contribution
AN - SCOPUS:85113531351
T3 - International Conference of Young Specialists on Micro/Nanotechnologies and Electron Devices, EDM
SP - 524
EP - 527
BT - 2021 IEEE 22nd International Conference of Young Professionals in Electron Devices and Materials, EDM 2021 - Proceedings
PB - IEEE Computer Society
T2 - 22nd IEEE International Conference of Young Professionals in Electron Devices and Materials, EDM 2021
Y2 - 30 June 2021 through 4 July 2021
ER -
ID: 34128734