NeMO-Net is a convolutional neural network (CNN) designed for marine ecosystem classification. The CNN takes as input 2D satellite and drone images as well as 3D reconstructions of underwater environments and generates classification maps for those environments as output. These classification maps can be used to better understand and protect coral reefs globally. One component of NeMO-Net is a citizen science game for mobile devices and personal computers. Through playing this game, players help NASA classify coral reefs and other aquatic ecosystems by painting on 2D and 3D images of coral. Players can rate the classifications of other players and level up in the food chain as they explore and classify coral reefs, other shallow marine environments, and creatures from locations all over the world. The application educates players on how to identify the different types of coral and player classifications are used to train the CNN to classify aquatic ecosystems autonomously.
Project URL: http://nemonet.info
Geographic Scope: Global - with focus on past and future field campaigns to Guam, Puerto Rico, Samoa, Marshall Islands, Red Sea, among other locations.
Project Status: Active - recruiting volunteers
Participation Tasks: Annotation, Data analysis, Data entry, Geolocation, Identification, Learning, Measurement, Observation, Problem solving, Sample analysis,
Start Date: 04/22/2020
Project Contact: firstname.lastname@example.org
Federal Government Sponsor:
Other Federal Government Sponsor:
Fields of Science: Animals, Archaeology and cultural, Biology, Climate and weather, Computers and technology, Disaster response, Ecology and environment, Education, Food, Geography, Geology and earth science, Health and medicine, Nature and outdoors, Ocean/water and marine, Science policy, Social science
Intended Outcomes: Map Earth's Oceans in this videogame that trains an AI for the NASA supercomputer using FluidCam's 3D images of the seafloor, the first instrument that can see through waves