Grumpy cat? AI app can decode feline feelings
Is your feline friend looking a bit grumpy?
Notoriously inscrutable - cats aren't the best a letting their owners know when something is wrong.
But a new app powered by artificial intelligence claims it can decode their subtle facial cues to deliver a speedy assessment of their well-being.
[MICHE PRIEST, VENTURE LEAD, SYLVESTER.AI]
"It helps human cat owners know if their cat is in pain or not. So it uses facial recognition technology. All you need to do is take a picture with your camera and then it can give you a result.”
The app - called Tably - uses the 'Feline Grimace Scale' or FGS pain assessment for cats based on: ear position, eye-narrowing, muzzle tension, whiskers change and head positioning.
Miche Priest is venture lead at the developers, Sylvester.ai.
"So the feline grimace scale, it is validated, there are papers that have been published on it, and it's a way for animal practitioners to be able to determine the level of pain that a cat is in based on those facial characteristics. And so we were able to train a machine using machine learning and a series of images to be able to make that determination. So if you're somebody who has a cat but you don't have that training, you can just use the app and it can help you determine how your cat is doing."
At Wild Rose Cat Clinic of Calgary where the developers trained the app's algorithm.
Dr. Liz Ruelle hopes it will aid novice cat owners and graduate vets.
[DR. LIZ RUELLE, WILD ROSE CAT CLINIC]
"I mean, what an amazing concept. When it was first discussed, I know I was so excited because; I love to say I'm good at what I do, I love working with cats, have always grown up with cats, I really am a crazy cat person first. So, I am quite adept at reading their body language. But for other colleagues, new grads, who maybe have not had quite so much experience, it can be very daunting to know is your patient painful? Are we doing a good job at what we do?"
However, the app is not quite perfect, say some early adopters.
Technology journalist Stephan Boissonneault tested the beta version, and hit a few snags.
[STEPHAN BOISSONNEAULT, WIRED UK CONTRIBUTOR]
"I used it when I was living with three cats. And one of the cats was black, and the machine, the app, felt it a little bit difficult to read the cat's face because black cats have darker faces, there's less light for the complete image. But it is machine learning so it will learn as more people use, that's the whole point of AI machine learning within an app."
