Improving Domain Generalization Via Event-Based Acoustic Scene Classification

Mezza AI, Sarti A. Improving Domain Generalization Via Event-Based Acoustic Scene Classification. In: AES Convention 153. Audio Engineering Society; 2022. Paper 5. Available from: https://aes.org/publications/elibrary-page/?id=21933

@inproceedings{Mezza2022_21933,
  author = {Mezza, Alessandro Ilic and Sarti, Augusto},
  title = {{Improving Domain Generalization Via Event-Based Acoustic Scene Classification}},
  booktitle = {AES Convention 153},
  note = {Paper 5},
  year = {2022},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=21933}
}

TY  - CPAPER
TI  - Improving Domain Generalization Via Event-Based Acoustic Scene Classification
AU  - Mezza, Alessandro Ilic
AU  - Sarti, Augusto
T2  - AES Convention 153
M1  - Paper 5
PY  - 2022
DA  - 2022/10/06
UR  - https://aes.org/publications/elibrary-page/?id=21933
PB  - Audio Engineering Society
LA  - en
AB  - Acoustic Scene Classification (ASC) has been typically addressed by feeding raw audio features to deep neural networks. However, such an audio-based approach has consistently proved to result in poor model generalization across different recording devices. In fact, device-specific transfer functions and nonlinear dynamic range compression highly affect spectro-temporal features, resulting in a deviation from the learned data distribution known as domain shift. In this paper, we present an alternative ASC paradigm that involves ditching the classic end-to-end audio-based training in favor of gathering an intermediate event-based representation of the acoustic scenes using large-scale pretrained models. Performance evaluation on the TAU Urban Acoustic Scenes 2020 Mobile Development dataset shows that the proposed event-based approach is up to 160% more robust than corresponding audio-based methods in the face of mismatched recording devices.
ER  -