Report of the Image Processing Workshop

Imagery and associated data collected from aircraft and vessel surveys have been critical for estimating marine mammal abundance, stock structure, behavior, health, and other information needed to meet the National Marine Fisheries Service’s (NOAA Fisheries) mission for decades. With the advent of digital photography, and most recently, advances in image resolution, unmanned aerial systems (UAS), autonomous underwater vehicles (AUV), and remote cameras, the total volume of electronic imagery collected by NOAA Fisheries researchers has increased rapidly. Biologists experience common challenges at regional NOAA Fisheries science centers, including the collection, storage, processing, and analysis of terabytes (TB) of data. For unmanned systems (UxS) to fully transition to operations in NOAA Fisheries, we must have efficient solutions for processing this unprecedented amount of image data. There is a strong need to increase efficiency by using a rtificial intelligence (AI) techniques, such as machine learning and computer vision, to automate image processing and analysis. The goals of the 2019 Image Processing Workshop were to bring NOAA Fisheries scientists, private industry, and external scientific collaborators together in order to understand how AI is being used to process marine mammal data in NOAA Fisheries, to discuss what is needed to harness this technology more effectively, to develop a community of NOAA Fisheries AI experts, and to train participants in available tools for implementing AI. 2020 NMFS (National Marine Fisheries Service) AFSC (Alaska Fisheries Science Center) Submitted https://doi.org/10.25923/7b2t-2g60 Public Domain 1862