Augmented reality as etiquette coach

Alasdair Allan has a practical goal for AR: putting names to faces.

Identifying local landmarks and uncovering hidden coupons are fun augmented reality applications, but “Programming iPhone Sensors” author Alasdair Allan has a loftier AR goal.

“I’m terrible with faces and names,” he said during a recent interview at OSCON. “So, I want those little glasses where you see someone and it’s like: ‘This is Gary. You met him in 2005. His wife is called Mary. He’s got three kids. His birthday is …’. That sort of thing. That’s my ultimate goal.”

Allan’s ideal is based on facial recognition, which is a step above facial detection. But you can’t have identification without detection, and detection is something we’re close to seeing in real-time. Allan himself successfully built a real-time face detection demo on the iPhone 3GS. The iPhone 4’s improved hardware makes the same functionality easier to implement. (Not trivial … just easier.)

Allan touched on a number of additional topics during our OSCON chat, including:

  • How (theoretically) a geolocation database — like SimpleGeo — could be matched up with a cloud-based facial recognition database.
  • How iPhoto’s Faces tool could influence FaceTime and other real-time video applications.
  • He closed with a brief rundown on the types of sensors commonly found in many smartphones … and why the introduction of gyroscopes in Android phones is now a near inevitability.

The full interview is available in the following video:


tags: , , ,
  • Joe McCarthy

    As another frequent sufferer of forgetting names and/or faces, this vision has always had some appeal, but I do have some reservations.

    I wonder how accurate the face recognition and inferencing would need to be for widespread adoption. Acting on one false positive – e.g., saying “Hi Gary” or “How’s Mary” when you are speaking to someone who is not Gary, and/or Mary is not his wife – may be enough for most users to abandon it.

    Also, whenever I’ve spoken with any of the wearable computing folks who use eyeglasses with displays (or retinal scanning), it’s always distracting to see their eye focus shift to read something – are they checking up on me or reading an email or tweet? Of course, many social conventions are undergoing change, and continuous partial attention / discrete distraction may become more widely socially acceptable in the future.

  • Tommy

    Hi guys,
    interesting post. Have you ever seen this multi-platform face recognition solution?

    Isn’t it the same kind of technology your talking about?



  • Mac Slocum

    @Tommy: Thanks for that link! It’s pretty close to what we were talking about. I’m guessing that solution is built around face *detection* rather than *recognition* though, right?

    Along those lines … I can only imagine the privacy debate that will rage when/if facial recognition is rolled out.