BROWSE
Detail
Conference Paper
Combining Multi-scale Features using Sample-level Deep Convolutional Neural Networks for Weakly Supervised Sound Event Detection
Cited - time in 
Share 
Share
- Authors
- Jongpil Lee, Jiyoung Park, Sangeun Kum, Youngho Jeong, Juhan Nam
- Issue Date
- 2017-11
- Citation
- Detection and Classification of Acoustic Scenes and Events (DCASE) 2017: Workshop, pp.69-73
- Language
- English
- Type
- Conference Paper
- 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 multiscaled 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 spectrogrambased 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.
- KSP Keywords
- Convolution neural network(CNN), Deep convolutional neural networks, Deep neural network(DNN), F-score, Multi-scale, Network method, Sound event detection(SED), Training samples, Weakly supervised, audio tagging, error rate