1

Apple’s FaceID on iPhone X works even in the dark using infrared light to “illuminate” your face with an IR projection and collect the features.

When designing the algorithms that confirm a user’s face, does Apple lower the “validation thresholds” when running in the dark to ensure a better user experience?

Can an attacker trying to fool FaceID with a replicated scanning target be more successful in poor lighting conditions?

schroeder
  • 123,438
  • 55
  • 284
  • 319
VAndrei
  • 127
  • 3
  • I don’t think it states explicitly that depth sense accuracy is not influenced by ambiental lighting. That’s the essence of the problem because heuristic algorithms are used to estimate the depth. I find it hard to believe that the algorithms perform the same no matter the lighting conditions. But perhaps this is not the right forum to ask this question. – VAndrei Dec 16 '19 at 21:35
  • That’s a weak assumption. Apple’s doc doesn’t state anyhting about how security of FaceID varies on external conditions. It’s just a marketing doc. https://www.extremetech.com/mobile/255771-apple-iphone-x-truedepth-camera-works this describes that flood illuminator is a technology to enhance usabilty in poor lighting so it looks like IR is not enough. – VAndrei Dec 16 '19 at 22:03
  • So fundamentally, a 3D printed mask would be able to fool a depth-based only system and that’s the reasoning for which iphone needs more than TrueDepth camera. And Flood Illuminator is a solution. Question is if it works similarly no matter the lighting. – VAndrei Dec 16 '19 at 22:10
  • 2
    So, now that you have some basic info, and you've provided foundational details, let's focus this question. – schroeder Dec 16 '19 at 22:34
  • If you looked up "Flood Illuminator" you will see that it, too, is IR. – schroeder Dec 16 '19 at 22:38

0 Answers0