A Robust Deep Learning-based System for Environmental Audio Compression and Classification

Emvoliadis A, Vryzas N, Stamatiadou ME, Vrysis L, Dimoulas C, Drosou A, Tzovaras D. A Robust Deep Learning-based System for Environmental Audio Compression and Classification. In: AES Convention 154. Audio Engineering Society; 2023. Paper 80. Available from: https://aes.org/publications/elibrary-page/?id=22105

@inproceedings{Emvoliadis2023_22105,
  author = {Emvoliadis, Alexandros and Vryzas, Nikolaos and Stamatiadou, Marina Eirini and Vrysis, Lazaros and Dimoulas, Charalampos and Drosou, Anastasios and Tzovaras, Dimitrios},
  title = {{A Robust Deep Learning-based System for Environmental Audio Compression and Classification}},
  booktitle = {AES Convention 154},
  note = {Paper 80},
  year = {2023},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=22105}
}

TY  - CPAPER
TI  - A Robust Deep Learning-based System for Environmental Audio Compression and Classification
AU  - Emvoliadis, Alexandros
AU  - Vryzas, Nikolaos
AU  - Stamatiadou, Marina Eirini
AU  - Vrysis, Lazaros
AU  - Dimoulas, Charalampos
AU  - Drosou, Anastasios
AU  - Tzovaras, Dimitrios
T2  - AES Convention 154
M1  - Paper 80
PY  - 2023
DA  - 2023/05/06
UR  - https://aes.org/publications/elibrary-page/?id=22105
PB  - Audio Engineering Society
LA  - en
AB  - In urban areas, the levels of air pollution play a significant role in the quality of life. These levels are rapidly increasing due to exponential urbanization. Although the field of real-time monitoring and forecasting of air pollution levels by fusing multiple modalities has been studied extensively, there is limited work on considering environmental audio in this fusion process. Moreover, studies on developed systems for noise pollution capturing and environmental sound classification have not investigated the potential compression of this considerable type of information. In the present work, a unified system for capturing and classifying urban environmental sounds is developed. Edge devices responsible for audio capturing (mobile phones, IoT nodes), compress the collected data and upload it to the Cloud. A server is then obliged for classifying this information to reveal its correlation with air pollution. We report a robust Auto-Encoder for Urban Environmental Audio Compression and Reconstruction, while the system’s performance on classification is very simillar to this of a State-of-the-Art classifier.
ER  -