Here’s some nerdy news: Israeli facial recognition startup Face.com has just opened up its API, allowing developers to integrate its facial recognition technology in third-party websites and applications. Since launching a year ago, the company has scanned more than 7 billion photos and tagged more than 52 million faces through its Photo Finder and Photo Tagger applications on Facebook.

Now, the technology is no longer limited to Facebook, as any third-party developer can integrate facial recognition into their own apps. The API uses a REST-like interface similar to Twitter’s API, and takes in URLs to photos.

I’m interested in seeing what kind of creative applications developers can come up with. They’ve post a few example apps already using the API, and there’s talk of upcoming facial recognition augmented reality apps.

If you’ve ever tried photographing a dog or cat, you probably know how difficult it can be to take a sharp photo while it’s looking at you. My friend’s dog (a pomeranian) is actually scared of my camera, and shies away when the DSLR is pointed at him.

FujiFilm’s new Finepix Z700 aims to make pet photographs easier by being the first camera to offer facial recognition for dogs and cats, and can automatically snap photographs for you when the pet is looking at the camera.

However, the technology is still pretty young, and has a ways to go before it rivals human facial recognition, which itself is ocassionally buggy.

For example, the camera has difficulty detecting pets that don’t stay still, and though it can detect up to 10 pet faces at once, it can’t handle a mix of dogs and cats. The subjects need to be either all dogs, or all cats.

Furthermore, some breeds of dogs (and maybe cats too?) can have pretty strange looking faces. The camera can’t handle those. FujiFilm even has a dedicated webpage listing the breeds of dogs and cats that the feature can usually detect, and includes sample images:

As you can see, you need to have a fairly… generic looking dog or cat if you want to detect its face.

Pets that cannot be easily detected include those that have: dark patches around the eyes or nose, too dark of a color, wrinkled/long/thin faces, or hair covering the eyes.

We’re guessing something like this will stump the camera:

Perhaps we should have titled this post, “Facial Recognition for Cute and Generic Looking Dogs and Cats”.

(via PC World)

Image credit: Castle Combe by Karen Roe