Can facial recognition software help bring endangered lemurs back from the brink?
The lemur situation in Madagascar is nothing short of a crisis. Ninety-four percent of the population is threatened with extinction, according to the International Union for Conservation of Nature, with habitat loss caused by the clearing of Madagascar's forests a big reason why. Researchers now have a new tool at their disposal, however, in the form of facial recognition software that can identify individual lemurs and inform long-term conservation strategies.
Found only on the island nation off Africa's south-east coast, lemurs are an iconic part of Madagascar's celebrated wildlife. But they now face a monumental fight to survive, with an estimated 80 percent of their forest habitat having been cleared away. This has left 24 of the 111 lemur species critically endangered. Forty-nine, meanwhile, are endangered while 20 other species are listed as vulnerable.
One way researchers can identify individual lemurs and track behavior and population numbers is by studying characteristics such as body shape and size, along with markers like injuries and scars. This is fine in the short term, but things can get complicated over longer periods as their physical appearance changes with time, making it difficult to tell one from the other.
Looking to overcome this dilemma, researchers at Michigan State University took the same facial recognition software used to find criminals and verify passports, and adapted it to the lemur world. This involved training the software using 462 images of 80 red-bellied lemurs and 190 images of other species to form the basis of the facial recognition system. Dubbed LemurFaceID, the team found that the software could correctly identify more than 100 individual lemurs at an accuracy of 98.7 percent.
"Like humans, lemurs have unique facial characteristics that can be recognized by this system," said Anil Jain, leader of the research team. "Once optimized, LemurFaceID can assist with long-term research of endangered species by providing a rapid, cost-effective and accurate method for identification."
This offers conservationists a handy new way of building data on how individual lemurs are living in the wild, helping them understand things like reproduction rates along with infant and juvenile mortality, all with a view to obtaining a clearer picture of population growth and decline.
These types of technologies are poised to play a bigger role in animal conservation. Earlier this month, for example, biologists developed software to identify individual sharks by analyzing the trailing patterns on their dorsal fins. For its part, the Michigan University team believes the LemurFaceID can be used to identify other endangered species with variable facial hair and skin patterns, such as bears, red pandas, raccoons and sloths.
"Facial recognition technology has the potential to help safeguard our society," says Jain. "Adapting it to help save endangered species is one of its most inspiring uses."
The team's research was published in the journal BMC Zoology.
Source: Michigan State University