Lion Facial Recognition Debuts in Africa
This article appeared in the July issue of Scientific American. Read the digital version here. Even the king of the jungle can't escape getting his picture taken these days. In June the Kenya-based Lion Guardians launched the Lion Identification Network of Collaborators (LINC). The database of lion profiles was built with the first facial-recognition software specifically designed to analyze the mugs of these big cats and distinguish them from one another. With LINC, the conservation organization and other wildlife researchers will have an easier way to monitor the beasts' whereabouts.
Their movements throughout Africa are poorly understood, and tracking efforts come with a host of difficulties: GPS transmitters are expensive, run out of batteries every one to three years, and can be fitted only when an animal is sedated. In addition, unlike leopards, cheetahs and tigers—whose spots and stripes make identification fairly easy—adult lions lack recognizable coat patterns.
Within the next few months about 1,000 lions will be added to LINC; the more photographs that are entered, the more accurate the software will become at identifying an individual. By keeping tabs on the cats' peregrinations, conservationists can better understand where lions find mates, water and prey, for example, as well as the nuanced changes to population dynamics caused by human expansion.
One need not get up close and personal to capture useful pictures. Shots snapped from up to 100 feet away will do the trick, says Stephanie Dolrenry, co-founder of Lion Guardians. Photo bombers and the most skittish lions alike typically turn to look at their pursuers before running away.