Computer vision uses algorithms to interpret photographs and video. Marine-life identification applies such methods to underwater imagery to detect, classify, or count organisms, often using reference images labeled by experts.
What Exists
NOAA documents machine-learning methods that use annotated underwater imagery to detect and classify reef fish. These systems are trained and tested for particular datasets and subjects. An identification display for snorkelers would need separate assessment for local species, viewpoint, color changes underwater, and image quality.
Questions to Investigate
- Can a small camera adequately capture what a snorkeler is looking at?
- What taxonomic databases and verified images cover a particular reef?
- How should the display convey uncertainty and visually similar species?
- What analysis can happen on the mask, on a boat, or elsewhere over a network?
Selected Sources
- NOAA Fisheries — Optics Technology: At the Heart of Marine Research.
- NOAA Ocean Exploration — Benthic Deepwater Animal Identification Guide (a scientific reference resource, not a snorkeling classifier).