Facial Recognition Tech Could Help Save Endangered Lemurs
A new system to identify the primates earns high scores for accuracy, offering the promise of better long-term tracking data to inform conservation strategies.
Could endangered lemurs in Madagascar benefit from the same technology used to identify criminals? New research into the prospects of using facial recognition on the distinctive primates suggests that signs point to yes.
For a study just published in the journal BMC Zoology, scientists from Michigan State, George Washington University and University of Arizona modified a human facial recognition system for use on lemurs, using more than 600 images of various species to amass a small dataset.
Lemurs make good candidates for such a system because they possess unique facial features such as fur patterns, said the researchers.
Dubbed LemurFaceID, the system was able to score 98.7 percent (+/- 1.8 percent) accuracy when put to the task of identifying more than 100 individual lemurs.
While the system is still a work in progress - field-ready facial recognition systems would use much larger datasets - the researchers say down the line the approach could help generate long-term data on lemurs, something sorely lacking. While short-term studies can get by identifying individual lemurs from characteristics such as body shape and size or distinctive injuries, such techniques would become cumbersome if applied to long-range studies of entire populations. As lemurs change over time, they're more difficult to uniquely identify using such methods.
Consequently, long-term lemur surveys are not very common. Now, though, data gaps about the primates could be filled.
"Studying lemur individuals and populations over long periods of time provides crucial data on how long individuals live in the wild, how frequently they reproduce, as well as rates of infant and juvenile mortality and ultimately population growth and decline," said study co-author Stacey Tecot, of University of Arizona, in a statement. "Using LemurFaceID can inform conservation strategies for lemurs, a highly endangered group of mammals."
What's more, the system could provide a less invasive way to identify individual creatures. Today researchers typically capture and place GPS collars on lemurs they want to be able to track, but that approach risks harming or stressing the animals. Too, veterinarians and anesthetics are necessary, raising the cost of the research process.
System designer Anil Jain, a biometrics expert and professor at Michigan State University, believes LemurFaceID can even be adapted to other types of animal, so long as they have variable facial hair and skin patterns. Bears, raccoons, sloths and red pandas are some of the animals that would fit that bill, Jain said, calling the effort to help save endangered species via facial recognition "one of its most inspiring uses."
"Our results suggest that human facial recognition techniques can be modified for identification of individual lemurs based on variation in facial patterns," the research team wrote. "Once optimized, our system can facilitate long-term research of known individuals by providing a rapid, cost-effective, and accurate method for individual identification."
WATCH VIDEO: How Exactly Do Our Brains Recognize Faces?