AI Training Data for Fisheries

More than a thousand commercial fishing boats carry electronic monitoring systems, which use on-board cameras to track catch and provide accountability in the global seafood market. These systems produce large volumes of video data that are screened by trained human reviewers. Over the next ten years, the number of boats carrying electronic monitoring systems is expected to grow by 10-20x, outpacing current review capacity.

We think AI can reduce the time needed for video review and lower the cost of electronic monitoring. Large open datasets like Imagenet, COCO and OIDV5 have supported the development of AI in other domains that can match or exceed human performance on visual recognition tasks. However, these sites contain little or no fisheries data.

We are partnering with governments, the fishing industry, electronic monitoring service providers, and others to create this open library of training data to accelerate the implementation of AI in fisheries. We are actively working with our partners to source and annotate images to add to the datasets.