TitleAnalytic framework for automatic event classification in sound streams from the Amazon forest
Professor/a organitzador/aDomènec Savi Puig Valls
InstitutionUniversitat Politècnica de Catalunya
Date 11-10-2017 13:00
SummaryThe recently started Project Providence aims at monitoring biodiversity of tropical forests with autonomous sensor nodes. Acoustic sensors will play a key role because they detect animal sounds over distances of up to several hundred meters. In this talk we first briefly present the characteristics of rainforest soundscapes. Then, we present our analytic system for automatic classification of acoustic events with focus on Machine Learning. Finally, we discuss how we plan to extend the predictive capabilities of the system.
Short Bio: Serge Zaugg has an academic training in statistics. In his early career, he worked in applied machine learning related to acoustic and radar sensors. In recent years he worked as data scientist in diverse fields such as biomedical imaging, clinical studies, public transportation, and radar technology. Currently he is at the Laboratory of Applied Bioacoustics (UPC) where he is responsible for data analytics.