DC Field | Value | Language |
---|---|---|
dc.contributor.author | Lee, Jongpil | ko |
dc.contributor.author | Park, Jiyoung | ko |
dc.contributor.author | Kum, Sangeun | ko |
dc.contributor.author | Jeong, Youngho | ko |
dc.contributor.author | Nam, Juhan | ko |
dc.date.accessioned | 2018-02-21T04:13:42Z | - |
dc.date.available | 2018-02-21T04:13:42Z | - |
dc.date.created | 2017-12-19 | - |
dc.date.issued | 2017-11-17 | - |
dc.identifier.citation | Proceedings of the 2nd Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE) | - |
dc.identifier.uri | http://hdl.handle.net/10203/239538 | - |
dc.description.abstract | This paper describes our method submitted to large-scale weakly supervised sound event detection for smart cars in the DCASE Challenge 2017. It is based on two deep neural network methods suggested for music auto-tagging. One is training sample-level Deep Convolutional Neural Networks (DCNN) using raw waveforms as a feature extractor. The other is aggregating features on multi-scaled models of the DCNNs and making final predictions from them. With this approach, we achieved the best results, 47.3% in F-score on subtask A (audio tagging) and 0.75 in error rate on subtask B (sound event detection) in the evaluation. These results show that the waveform-based models can be comparable to spectrogram-based models when compared to other DCASE Task 4 submissions. Finally, we visualize hierarchically learned filters from the challenge dataset in each layer of the waveform-based model to explain how they discriminate the events. | - |
dc.language | English | - |
dc.publisher | IEEE | - |
dc.title | Combining Multi-Scale Features Using Sample-level Deep Convolutional Neural Networks for Weakly Supervised Sound Event Detection | - |
dc.type | Conference | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | Proceedings of the 2nd Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE) | - |
dc.identifier.conferencecountry | GE | - |
dc.identifier.conferencelocation | Munich | - |
dc.contributor.localauthor | Nam, Juhan | - |
dc.contributor.nonIdAuthor | Park, Jiyoung | - |
dc.contributor.nonIdAuthor | Jeong, Youngho | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.