Dataset Retrieval for the 2015 DCLDE Workshop

As in previous workshops, a common dataset of underwater recordings is being made available to encourage researchers to compare results, and find original solutions to our shared problems. Participants are encouraged to present the results of their work at the workshop. For the 2015 workshop, multiple datasets have been created to address some of the gaps identified in discussions at previous workshops. In particular, we were interested in providing large training and testing datasets that cover a range of spatial, temporal, and recording variability that is encountered by researchers in the field. We hope that these datasets will provide an opportunity to develop detectors and classifiers that will perform more robustly across different tasks and in new conditions.

We will provide two different datasets, one focusing on high-frequency (up to 160 kHz) and one on low-frequency (up to 1.6 kHz) bandwidth signals. The high-frequency dataset will consist of marked encounters with echolocation clicks of species commonly found along the US West Coast, including Cuvier’s and Baird’s beaked whales, Risso’s and Pacific white-sided dolphins, sperm whales, unidentified porpoises and unidentified odontocetes. The low-frequency dataset will contain call-level markings for blue whale D calls and fin whale 40 Hz calls. In both cases, the recordings were collected with High-frequency Acoustic Recordings Packages (HARPs) deployed off the southern and central coast of California at different locations, spanning all four seasons. The workshop datasets have been provided by Ana Širović and Simone Baumann-Pickering of the Scripps Institution of Oceanography. If you would like to learn more about these datasets, please check out the Dataset Documentation.