Deep Lake Explorer: Using citizen science to analyze underwater video from the Great Lakes
Citation:
Wick, M., M. Nord, M. Pawlowski, R. Debbout, J. Launspach, Dave Bolgrien, AND T. Angradi. Deep Lake Explorer: Using citizen science to analyze underwater video from the Great Lakes. AIGLR, Toronto, ON, CANADA, June 18 - 22, 2018.
Impact/Purpose:
Deep Lake Explorer: Using citizen science to analyze underwater video from the Great Lakes
Description:
While underwater video collection technology continues to improve, advancements in underwater video analysis techniques have lagged. Crowdsourcing image interpretation using the Zooniverse platform has proven successful for many projects, but few projects to date have included videos. Deep Lake Explorer is a web application hosted on the Zooniverse platform that allows citizens to interpret underwater video collected in the Great Lakes. Water clarity, lighting, camera movement, orientation, and naturally complicated ecosystems make underwater video interpretation subjective and challenging. Deep Lake Explorer facilitates the consensus of multiple users on the content of underwater videos, building precision. Users characterize habitat and identify presence/absence of round gobies and dreissenids in the Great Lakes nearshore. Poor quality videos are identified based on lack of consensus rather than interpretation by one analyst. An initial beta-testing phase allows users to provide input on design, support materials, and analysis workflows. During beta-testing, researchers also determine the number of views required to establish consensus, based on the dataset and analysis type. More views are necessary if the analysis is difficult. Deep Lake Explorer is facilitating the analysis of 729 videos, with 54 included in beta-testing. Results from the application will inform habitat and fisheries management programs.