GTM participates in the Urban Sound Symposium 2021
Last 19th-21st of April, two works from the GTM research group were presented at the second edition of the Urban Sound Symposium. This year, the conference was organized simultaneously in Ghent, Montreal, Nantes, Zurich, London and Berlin. Due to COVID-19 pandemic, however, the conference was held on a virtual modality. Despites of the virtual conditions, the congress was a great opportunity for researchers from La Salle to interact with worldwide investigators in poster-booths, virtual coffee tables, and even in social activities.
The funny and intuitive virtual framework allowed a direct communication among all the attendees, and the plenary talks by the eminent speakers led to interesting discussion sessions.
The two works from the group presented at the conference were:
- "Sons al Balcó": Soundscape of Catalonia during the first COVID-19 lockdown, by the authors Xavier Baño, Pau Bergadà, Alba Egea, Maria Foraster, Marc Freixes, Gerardo José Ginovart-Panisello, Roger Mallol, Xavier Martín, Anna Martíez, Carme Martínez-Suquía, Ferran Orga, Xavier Sevillano and Rosa Ma Alsina Pagès, an interdisciplinary project from La Salle - URL and ISGlobal.
Dr. Rosa Ma Alsina Pagès presented this poster, in which she explained and summarized the outcomes of the research project “Sons al Balcó”, partially funded by the Ramon Llull University. The idea behind this project is to characterize the soundscape of Catalonia by analyzing the audiovisual content of a set of video recordings uploaded to a web server by volunteer citizens during the 2020 lockdown. In order to go further with the soundscape of Catalonia analysis, in 2021 the project has developed another full collecting campaign.
2. Prototyping a low-cost wireless acoustic sensor network with physical redundancy to automatically classify acoustic events in urban environments, by the authors Ester Vidaña-Vila, Rosa Ma Alsina-Pagès and Joan Navarro.
Ester Vidaña-Vila, a PhD candidate from the group, presented this work, that summarizes a part of her PhD thesis. Concretely, this work conceives the prototype of a wireless acoustic sensor network aimed to detect acoustic urban events (e.g. car horns, sirens, people talking, etc.). The work describes the proposed hardware architecture, the deep-learning based software running on each individual node and discusses the usage of physical redundancy of sensors to improve the confidence and accuracy of the system.