MENU

Tracking Lemurs Could Become Less Invasive, Thanks To Facial Recognition Software

Close

Facial recognition systems have long been considered helpful tools in saving endangered animals, including North Atlantic right whales.

Now, in a new study, a team of scientists in the United States investigated whether such software could also benefit ongoing research on lemurs, and ultimately, save the species from harm.

Tracking Lemurs In The Wild

More than 100 species of lemur are known to be living throughout Madagascar. In 2014, a report revealed that 90 percent of these lemur species are on the verge of extinction because of a dramatic increase in illegal poaching.

In the new research, biological anthropologist Rachel Jacobs of George Washington University and her colleagues successfully developed a facial recognition software called LemurFaceID, which can track and identify individual lemurs in the wild in a less invasive manner.

LemurFaceID is designed to identify an individual lemur based on photographs. It's very convenient because it can help researchers build an accurate database of lemurs in Madagascar over time and aid in conservation efforts.

Jacobs explained that she and senior study author Stacey Tecot of University of Arizona were not satisfied with conventional approaches used in lemur studies.

Past research required long-term life history data on lemurs in the aspects of reproduction, survival, and population growth. These required scientists to trap and physically tag the animals, which could potentially harm them and put them under stress.

"We aimed to do something different with red-bellied lemurs," said Jacobs. "We sought the expertise of our computer science collaborators."

Indeed, LemurFaceID is a non-invasive, cost-effective, faster, and more accurate way of tracking lemurs. It allows scientists to upload photos of the animals to the system and save the data.

Animal Conservation Efforts

Anil Jain, one of the researchers and a biometrics expert at Michigan State University, said lemurs' facial features can be easily recognized by the software, which has a 98.7 percent accuracy rate.

To test whether the software would work, Jain and his colleagues used around 462 images of 80 red-bellied lemurs and another 190 images of other lemur species. Many of these images were captured in Ranomafana National Park in Madagascar.

In the end, Jain and his teammates believe the software not only works on lemurs, but on other animals as well.

They believe it could also assist in other forms of animal conservation, especially among primate and non-primate species with variable facial hair and skin patterns. These include bears, raccoons, sloths, and red pandas.

"Adapting it to help save endangered species is one of its most inspiring uses," added Jain.

The findings of the study are issued in the journal BMC Zoology.

ⓒ 2018 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Real Time Analytics