Frequency Dependent Sound Event Detection for DCASE 2022 Challenge Task 4
Abstract
While many deep learning methods on other domains have been applied to sound event detection (SED), differences between original domains of the methods and SED have not been appropriately considered so far. As SED uses audio data with two dimensions (time and frequency) for input, thorough comprehension on these two dimensions is essential for application of methods from other domains on SED. Previous works proved that methods those address on frequency dimension are especially powerful in SED. By applying FilterAugment and frequency dynamic convolution those are frequency dependent methods proposed to enhance SED performance, our submitted models achieved best PSDS1 of 0.4704 and best PSDS2 of 0.8224.
Keywords
Cite
@article{arxiv.2206.11645,
title = {Frequency Dependent Sound Event Detection for DCASE 2022 Challenge Task 4},
author = {Hyeonuk Nam and Seong-Hu Kim and Deokki Min and Byeong-Yun Ko and Seung-Deok Choi and Yong-Hwa Park},
journal= {arXiv preprint arXiv:2206.11645},
year = {2022}
}
Comments
Technical Reprot submitted for DCASE2022 Challenge Task4