Tampere University of Technology

TUTCRIS Research Portal

Direction of Arrival Estimation for Multiple Sound Sources Using Convolutional Recurrent Neural Network

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Standard

Direction of Arrival Estimation for Multiple Sound Sources Using Convolutional Recurrent Neural Network. / Adavanne, Sharath; Politis, Archontis; Virtanen, Tuomas.

2018 26th European Signal Processing Conference (EUSIPCO). IEEE, 2018. p. 1462-1466.

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Harvard

Adavanne, S, Politis, A & Virtanen, T 2018, Direction of Arrival Estimation for Multiple Sound Sources Using Convolutional Recurrent Neural Network. in 2018 26th European Signal Processing Conference (EUSIPCO). IEEE, pp. 1462-1466, European Signal Processing Conference, 1/01/00. https://doi.org/10.23919/EUSIPCO.2018.8553182

APA

Vancouver

Author

Adavanne, Sharath ; Politis, Archontis ; Virtanen, Tuomas. / Direction of Arrival Estimation for Multiple Sound Sources Using Convolutional Recurrent Neural Network. 2018 26th European Signal Processing Conference (EUSIPCO). IEEE, 2018. pp. 1462-1466

Bibtex - Download

@inproceedings{4eb0a5929ee442bb8d208403b2a8d680,
title = "Direction of Arrival Estimation for Multiple Sound Sources Using Convolutional Recurrent Neural Network",
abstract = "This paper proposes a deep neural network for estimating the directions of arrival (DOA) of multiple sound sources. The proposed stacked convolutional and recurrent neural network (DOAnet) generates a spatial pseudo-spectrum (SPS) along with the DOA estimates in both azimuth and elevation. We avoid any explicit feature extraction step by using the magnitudes and phases of the spectrograms of all the channels as input to the network. The proposed DOAnet is evaluated by estimating the DOAs of multiple concurrently present sources in anechoic, matched and unmatched reverberant conditions. The results show that the proposed DOAnet is capable of estimating the number of sources and their respective DOAs with good precision and generate SPS with high signal-to-noise ratio.",
keywords = "array signal processing, direction-of-arrival estimation, feature extraction, feedforward neural nets, recurrent neural nets, signal classification, spatial pseudospectrum, SPS, DOA estimates, explicit feature extraction step, DOAnet, multiple concurrently present sources, anechoic unmatched reverberant conditions, matched unmatched reverberant conditions, arrival estimation, multiple sound sources, convolutional recurrent neural network, deep neural network, Direction-of-arrival estimation, Estimation, Azimuth, Feature extraction, Spectrogram, Multiple signal classification, Two dimensional displays",
author = "Sharath Adavanne and Archontis Politis and Tuomas Virtanen",
note = "EXT={"}Politis, Archontis{"}",
year = "2018",
month = "9",
doi = "10.23919/EUSIPCO.2018.8553182",
language = "English",
isbn = "978-1-5386-3736-4",
publisher = "IEEE",
pages = "1462--1466",
booktitle = "2018 26th European Signal Processing Conference (EUSIPCO)",

}

RIS (suitable for import to EndNote) - Download

TY - GEN

T1 - Direction of Arrival Estimation for Multiple Sound Sources Using Convolutional Recurrent Neural Network

AU - Adavanne, Sharath

AU - Politis, Archontis

AU - Virtanen, Tuomas

N1 - EXT="Politis, Archontis"

PY - 2018/9

Y1 - 2018/9

N2 - This paper proposes a deep neural network for estimating the directions of arrival (DOA) of multiple sound sources. The proposed stacked convolutional and recurrent neural network (DOAnet) generates a spatial pseudo-spectrum (SPS) along with the DOA estimates in both azimuth and elevation. We avoid any explicit feature extraction step by using the magnitudes and phases of the spectrograms of all the channels as input to the network. The proposed DOAnet is evaluated by estimating the DOAs of multiple concurrently present sources in anechoic, matched and unmatched reverberant conditions. The results show that the proposed DOAnet is capable of estimating the number of sources and their respective DOAs with good precision and generate SPS with high signal-to-noise ratio.

AB - This paper proposes a deep neural network for estimating the directions of arrival (DOA) of multiple sound sources. The proposed stacked convolutional and recurrent neural network (DOAnet) generates a spatial pseudo-spectrum (SPS) along with the DOA estimates in both azimuth and elevation. We avoid any explicit feature extraction step by using the magnitudes and phases of the spectrograms of all the channels as input to the network. The proposed DOAnet is evaluated by estimating the DOAs of multiple concurrently present sources in anechoic, matched and unmatched reverberant conditions. The results show that the proposed DOAnet is capable of estimating the number of sources and their respective DOAs with good precision and generate SPS with high signal-to-noise ratio.

KW - array signal processing

KW - direction-of-arrival estimation

KW - feature extraction

KW - feedforward neural nets

KW - recurrent neural nets

KW - signal classification

KW - spatial pseudospectrum

KW - SPS

KW - DOA estimates

KW - explicit feature extraction step

KW - DOAnet

KW - multiple concurrently present sources

KW - anechoic unmatched reverberant conditions

KW - matched unmatched reverberant conditions

KW - arrival estimation

KW - multiple sound sources

KW - convolutional recurrent neural network

KW - deep neural network

KW - Direction-of-arrival estimation

KW - Estimation

KW - Azimuth

KW - Feature extraction

KW - Spectrogram

KW - Multiple signal classification

KW - Two dimensional displays

U2 - 10.23919/EUSIPCO.2018.8553182

DO - 10.23919/EUSIPCO.2018.8553182

M3 - Conference contribution

SN - 978-1-5386-3736-4

SP - 1462

EP - 1466

BT - 2018 26th European Signal Processing Conference (EUSIPCO)

PB - IEEE

ER -