Speaker: Sergio Segovia González. Abstract: Traditional SED approaches are based on either specialized models or on these models in combination with general audio embedding extractors. In this article we propose to reframe SED as an object detection task in the… Read More
