Algorithm Research and Modeling
Over the past two decades, BRP has collected large amounts — literally, hundreds of years — of acoustic data. To keep up with the analysis of the data, we develop detection, classification, and localization (DCL) algorithms to efficiently and semi-automatically screen our sound archives for signals of interest. Recently we have started to evaluate machine learning (especially deep learning) approaches to improve the performance of our automated analysis approaches. Newly developed algorithms are made available to the research community through our Raven Pro and Raven-X software packages.
We are also working on the development of complex models to assess the impact of anthropogenic noise on the communication space of animals as well as visualization tools to convey our research results to the public and policy makers.