Finding Lost Pets With Facial Recognition Technology


How do you get a dog to pose for a photo? Bribe them with treats? Wave pictures of cats at them? Do a funny dance? John Polimeno, the CEO of Finding Rover, says the answer is to play them squealing puppy sounds. This makes pooches freeze long enough to capture their face using the pet-recognition software in his app. Earlier attempts involved crying baby sounds and squeaky toys — but pup whining proved most effective.

Polimeno was inspired to solve the problem of lost dogs after spotting “missing” signs in a coffee shop. He thought there had to be a more efficient method than posters — and wondered if human facial-recognition technology could work with animals. Lacking a tech background, he connected with software developers at the University of Utah, commissioning a study. “A person is easier to identify than an animal — our noses are in the same spot, and our chin and eyes,” he says. They developed an algorithm that he says can identify dogs with 98 percent accuracy. He emphasized the notion of “crowd-finding,” in which the public all contribute to connecting lost pups with their owners.

People register their pets on the app by adding details like name, breed, sex and age — and the all-important photo. Ready to take your pooch’s snap? Just press the “bark button,” which plays the whining sound before the camera shutter goes off. If your dog goes missing, you can issue alerts with contact information. Shelters and vets that have signed up are also alerted, maximizing the chance of recovery (110 across the world have registered). Launched on the Apple store in 2013, and online and for Android in 2014, the app has more than 100,000 dogs registered.

The next step is cats — and Finding Rover is planning to introduce Finding Kitty within two months. Because cats’ faces have more in common, the recognition accuracy is 99 percent — so “maybe Rover will find a friend named Fluffy,” Polimeno says.

Click here to read more.

SOURCE: Ozy, Zara Stone