Face Recognition Technology for Turtles
It’s interesting to see how new technologies are deployed and adopted over time. Many of them, including face recognition, have controversial use cases when they are first introduced. However, a natural part of the technology adoption cycle is creating written rules (laws) and unwritten rules (customs) that govern how a new technology is used — what we want (less waiting in lines, unlocking phones) and what we don’t want (mass surveillance). It's important that diverse groups of people build the AI systems of the future so that they are equitable, fair and transparent.
Face recognition technology used to identify turtles is a great example of a new and unexpected use case. Turtles are considered by biologists to be an indicator species. An indicator species is a species whose behavior helps scientist understand the underlying welfare of the entire ecosystem. For example, the presence of otters in rivers has been considered a sign of a clean, healthy river, since a ban on chlorine pesticides in the 1970s brought the species back from the brink of extinction. A thriving population of otters is an indication of a healthy ecosystem.
Turtles cultivate their ecosystem by grazing on seagrass, which provides a habitat for numerous fish and crustaceans. Traditionally, individual turtles have been identified and tracked by biologists with physical tags. This method has many problems including the frequent loss or erosion of these tags in seawater. Additionally, it’s challenging to keep a turtle still enough to locate and record the tag.
So researchers are solving these problems with turtle facial recognition. This works because, like people, turtles have a unique pattern around their face and this pattern is unique (enough) and remains unchanged over their multi-decade lifespan. This technology increases the reliability and speed of turtle re-identification, and potentially offers a way to replace the use of uncomfortable physical tags altogether. Similar efforts are underway to identify farm livestock as well.
Just like in human face recognition technology, researchers collected images of turtles from different angles and used them to train an algorithm. Also like in human face recognition technology, differences in lighting, pose, scale and background add complexity to the project.
When the CIA invented face recognition, they probably never envisioned its use for turtles and conservation. But artificial intelligence (AI) systems are making progress in conservation efforts worldwide, helping tackle complex problems at scale – from studying the behavior of animal communities in the Serengeti to help conserve the diminishing ecosystem, to spotting poachers and their wounded prey to prevent species going extinct.
