advertisement
advertisement
  • 03.27.15

Tag A Tiger: How Facial Recognition Can Help Track Endangered Animals

Have you seen this cat?

Tag A Tiger: How Facial Recognition Can Help Track Endangered Animals
[Top Photo: ehtesham via Shutterstock]

When researchers try to count the world’s quickly disappearing population of wild tigers–only about 3,000 are left, at last estimate–the process is usually slow and complicated. Biologists might spend months following tracks or setting up camera traps to snap photos as tigers walk by. In some countries, tigers are poached faster than they can be counted.

advertisement

A new iPad app takes a different approach. By mining the thousands of tiger photos that are already online, the project will eventually use facial recognition to identify individual tigers and track the global population in real time.


“There’s enough data online through people’s photographs,” says Aaron Mason, a PhD student at the UK’s University of Surrey, who helped create the app. “If we can just analyze this data, we can find out a lot more than what individuals are working on all around the world. It becomes easier to keep on top of things. Rather than doing a survey every few years, you could literally have a look at how many tigers there are in the world this hour, and what’s changed in the last day.”

The app, called Wildsense Tigers, pulls tourist photos from sites like Flickr and Instagram, along with photos from biologists’ official camera traps in the wild.


“Tigers are quite elusive, and it would be quite rare to bump into one–and to then at that moment want to take a photo of it, rather than running in the opposite direction. So tourist photographs are just one source,” says Mason. “We also have camera track footage from certain areas, as well as photos from the experts whose job is to track these elusive animals.”

To sort through all the photos, the researchers are turning to anyone with an iPad. The app is set up like a game: First, someone identifies that she’s actually looking at a real tiger, and then draws a box around the tiger’s face and tags it with location or any other data that can be gleaned from the photo. As all the data is collected, the researchers will use it to finish building an algorithm that will automatically recognize individual tigers.

“Each tiger has a unique stripe pattern–it’s like a fingerprint or a barcode,” says Mason. “An expert is able to look at a tiger and instantly know which one it is. So it’s possible, but it is a challenge, and it’s especially a challenge for various photographs where there’s a difference in light or occlusions. It’s quite ambitious, really, what we’re trying to do.”

advertisement

Despite the challenges, the researchers already have a prototype that worked well with a small number of tigers. Now, they’re hoping to use it on a global scale, and eventually use a similar program to track other wild animals. (They also plan to develop the software to recognize things like brands or objects, as a way to monetize it outside of conservation.)

“What I find inspiring about this is there’s a massive amount of information on the Internet, and all you have to do is bring it together, and you have knowledge which never existed before,” says Mason. “And there are people around the world who want to help–if you give them a platform, you can do a lot of things for good.”

Download Wildsense Tigers at the App Store.

About the author

Adele Peters is a staff writer at Fast Company who focuses on solutions to some of the world's largest problems, from climate change to homelessness. Previously, she worked with GOOD, BioLite, and the Sustainable Products and Solutions program at UC Berkeley.

More

Video