Wild Lemurs Get Their Own Facial Recognition Software

Do these lemurs look identical to you? New face recognition software can tell them apart. Bernard Castelein/Nature Picture Library/Getty Images

Goodbye chunky ear tags and radio collars, not to mention dart guns and tranquilizers. Soon it may be possible to use non-invasive computer-assisted facial recognition to study and track endangered mammals in the wild, according to results of a study just published in BMC Zoology.

Researchers first developed a computer-assisted facial recognition system to identify individual lemurs in the wild based on their mug shots. The system, LemurFaceID, is able to do so, in part, by ignoring the effects of facial hair and ambient lighting during the identification process. The scientists tested their system using pix of wild, red-bellied lemurs from Madagascar's Ranomafana National Park. After 100 trials, the results showed LemurFaceID had a 98.7 percent success rate (plus or minus 1.81 percent).

Lemurs live in the trees of Madagascar, an island off the coast of Africa. The primates resemble a cross between a squirrel and a monkey, and in 2012 were named the most endangered mammal in the world. The new face-recognition system has great potential to help the primates in several ways.

First, LemurFaceID will help researchers develop a database of Madagascar's red-bellied lemurs and study them over long periods of time, providing valuable insights into reproduction, population growth and survival. The system may even be able to help bust lemur-traffickers. In addition, LemurFaceID may negate the need to trap and physically tag lemurs in order to track them. The system is also fast and cost-effective.

This image shows the similarities and differences between the 80 lemurs that were photographed for this database.
George Washington University/Creative Commons Attribution 4.0 International License

While LemurFaceID was created specifically for lemurs, scientists say the system has the potential to assist in the study and conservation of other species that have identifiable variations in their hair and skin patterns, such as red pandas and some bears. This could do away with the need for tagging these animals for tracking purposes.