There is presently a limited number of computer vision capabilities in Apple’s impressive technology stack. The main one that comes to mind is the face detection functionality. You can use it in the Camera app on your iDevice and in the Photos app on your Mac.
As you may know, I am mostly interested in 3D computer vision. I consider face recognition to be 2D trickery. In the huge number of Apple frameworks I have found only one capability that I would categorise as 3D vision. It can actually detect the orientation-in-depth of a face – the amount that the head is rotated around the vertical axis. You can find it in the AVFoundation framework as the read only property yawAngle of AVMetadataFaceObject. It can only detect frontal (0 degrees), three-quarter (45 or 315 degrees), or side views (90 or 270 degrees).
Over the last few years, Apple acquired several companies that were specialised in computer vision and augmented reality. For example, PrimeSense was bought in 2013 and was famous for providing the visual depth detection technology for the first generation of Microsoft’s Xbox Kinect systems. Another example is Metaio that was bought last year. That company was popular for its augmented reality SDK. Metaio cancelled all products and services after the acquisition.
So, Apple now has quite a lot of computer vision knowledge in-house. With WWDC 2016 starting in about an hour, I feel it is my duty to report on all new 3D computer vision announcements.
My sense is that computer vision technology has not matured enough to live up to Apple’s high standards. So my prediction is that there will be very little to report.
But I hope I am wrong.
> You can sign up for my newsletter here.
Stijn Oomes PhD 