Evangelos Kazakos1, Arsha Nagrani2, Andrew Zisserman2 and Dima Damen1

1University of Bristol, Dept. of Computer Science, 2University of Oxford, VGG

Overview

Abstract

We propose a two-stream convolutional network for audio recognition, that operates on time-frequency spectrogram inputs. Following similar success in visual recognition, we learn Slow-Fast auditory streams with separable convolutions and multi-level lateral connections. The Slow pathway has high channel capacity while the Fast pathway operates at a fine-grained temporal resolution. We showcase the importance of our two-stream proposal on two diverse datasets: VGG-Sound and EPIC-KITCHENS-100, and achieve state-of-the-art results on both.

Downloads

Bibtex

@ARTICLE{Kazakos2021SlowFastAuditory,
   title={Slow-Fast Auditory Streams For Audio Recognition},
   author={Kazakos, Evangelos and Nagrani, Arsha and Zisserman, Andrew and Damen, Dima},
           journal   = {CoRR},
           volume    = {abs/2103.03516},
           year      = {2021},
           ee        = {https://arxiv.org/abs/2103.03516},
}

Acknowledgements

Kazakos is supported by EPSRC DTP, Damen by EPSRC Fellowship UMPIRE (EP/T004991/1) and Nagrani by Google PhD fellowship. Research is also supported by Seebibyte (EP/M013774/1).