Detail publikačního výsledku

Convolutional Neural Networks and X-Vector Embedding for DCASE2018 Acoustic Scene Classification Challenge

ZEINALI, H.; BURGET, L.; ČERNOCKÝ, J.

Originální název

Convolutional Neural Networks and X-Vector Embedding for DCASE2018 Acoustic Scene Classification Challenge

Anglický název

Convolutional Neural Networks and X-Vector Embedding for DCASE2018 Acoustic Scene Classification Challenge

Druh

Stať ve sborníku mimo WoS a Scopus

Originální abstrakt

In this paper, the Brno University of Technology (BUT) team submissionsfor Task 1 (Acoustic Scene Classification, ASC) of theDCASE-2018 challenge are described. Also, the analysis of differentmethods on the leaderboard set is provided. The proposedapproach is a fusion of two different Convolutional Neural Network(CNN) topologies. The first one is the common two-dimensionalCNNs which is mainly used in image classification. The second oneis a one-dimensional CNN for extracting fixed-length audio segmentembeddings, so called x-vectors, which has also been used inspeech processing, especially for speaker recognition. In additionto the different topologies, two types of features were tested: logmel-spectrogram and CQT features. Finally, the outputs of differentsystems are fused using a simple output averaging in the bestperforming system. Our submissions ranked third among 24 teamsin the ASC sub-task A (task 1a).

Anglický abstrakt

In this paper, the Brno University of Technology (BUT) team submissionsfor Task 1 (Acoustic Scene Classification, ASC) of theDCASE-2018 challenge are described. Also, the analysis of differentmethods on the leaderboard set is provided. The proposedapproach is a fusion of two different Convolutional Neural Network(CNN) topologies. The first one is the common two-dimensionalCNNs which is mainly used in image classification. The second oneis a one-dimensional CNN for extracting fixed-length audio segmentembeddings, so called x-vectors, which has also been used inspeech processing, especially for speaker recognition. In additionto the different topologies, two types of features were tested: logmel-spectrogram and CQT features. Finally, the outputs of differentsystems are fused using a simple output averaging in the bestperforming system. Our submissions ranked third among 24 teamsin the ASC sub-task A (task 1a).

Klíčová slova

Audio scene classification, Convolutional neuralnetworks, Deep learning, x-vectors, Regularized LDA

Klíčová slova v angličtině

Audio scene classification, Convolutional neuralnetworks, Deep learning, x-vectors, Regularized LDA

Autoři

ZEINALI, H.; BURGET, L.; ČERNOCKÝ, J.

Rok RIV

2019

Vydáno

19.11.2018

Nakladatel

Tampere University of Technology

Místo

Surrey

ISBN

978-952-15-4262-6

Kniha

Proceedings of DCASE 2018 Workshop

Strany od

1

Strany do

5

Strany počet

5

URL

http://dcase.community/documents/workshop2018/proceedings/DCASE2018Workshop_Zeinali_149.pdf

BibTex

@inproceedings{BUT155111,
  author="Hossein {Zeinali} and Lukáš {Burget} and Jan {Černocký}",
  title="Convolutional Neural Networks and X-Vector Embedding for DCASE2018 Acoustic Scene Classification Challenge",
  booktitle="Proceedings of DCASE 2018 Workshop",
  year="2018",
  pages="1--5",
  publisher="Tampere University of Technology",
  address="Surrey",
  isbn="978-952-15-4262-6",
  url="http://dcase.community/documents/workshop2018/proceedings/DCASE2018Workshop_Zeinali_149.pdf"
}

Dokumenty

zeinali_dcase2018_149