I've been working on a team trying to tackle the problem of how to analyze the deluge of underwater video being collected. Why collect underwater video you ask? The reasons range from trying to answer foundational science questions, fisheries stock assessments, environmental surveys, etc. Specific examples include: How will deep sea mining affect the oceans? How will that offshore wind farm impact local ecosystem?

One tool that we've developed is #FathomNet (https://fathomnet.org)

/1

FathomNet

#FathomNet is a resource to help foster collaboration between the #MachineLearning and the #OceanScience communities.

We're working hard to aggregate expertly curated imaged data sets, sort of like ImageNet, but for the ocean. Our goal is to have amazing data sets that can be used by experts in #MachineLearning to help us, folks who are really good at identifying things in the ocean, but maybe less good at developing novel machine learning models, solve important science problems.

/2

If you're interested in contributing to this effort, either because you have expertise in #MachineLearning or maybe you're an deep-sea taxonomist or even if you're an educator and need data to help train the next generation of researchers, we're having a workshop in February.

The workshop will be February 22-23 from 1500-1900 PST (UTC-8). Please register at http://tinyurl.com/fathomnetws.

/3 fini

Welcome! You are invited to join a meeting: FathomNet Workshop. After registering, you will receive a confirmation email about joining the meeting.

FathomNet is an open-source image database that can be used to train, test, and validate state-of-the-art artificial intelligence algorithms to help us understand our ocean and its inhabitants (see recent article - https://www.nature.com/articles/s41598-022-19939-2) for more information. Join us for a 2-day workshop, including how you can contribute to and benefit from the FathomNet ecosystem, as well as how to get involved in the FathomNet community. When: February 22 and 23, 2023. 1500-1900 Pacific Time Registration is limited to 500 people. For more information: Web: http://fathomnet.org/ GitHub: https://github.com/fathomnet Medium: https://medium.com/fathomnet

Zoom
@Schlining Please tell me some of the machines need validation datasets that can be crowd sourced.
I miss the projects on #zooniverse that let me click on scallops and brittle stars and such. It was a fantastic break for my brain and I learned a lot about what the seafloor can look like.
@DrKylara Good news! We're actually working on developing a game for ocean enthusiasts to be involved and contribute. (Maybe no scallops though, but lots of brittlestars 🙂 ) That effort is just getting started, so that will likely be out later in 2023. Details at https://www.mbari.org/news/ocean-vision-ai-uses-the-power-of-artificial-intelligence-to-process-ocean-imagery/
Ocean Vision AI uses the power of artificial intelligence to process ocean imagery • MBARI

With funding from the National Science Foundation, MBARI and our collaborators will develop a machine-learning platform to process ocean video and imagery.

MBARI