Inpainting, Deep Learning, and Anonymization
AnonymizationImage recognition is one of the areas that Deep Learning excels at. It is, in fact, so good at it that these days you can pretty easily identify anyone from a photograph (especially if you’ve got a large enough corpus of identity information, like Facebook, Google, or, oh, the Feds).

Blurring: Not the AnswerWell, you guessed wrong! Back in 2016, McPherson et al. (•) showed that you can uniquely identify people from these types of blurred images. (You’ve got to remember, Neural Networks don’t give a s**t about whether a thing is a “face” or not, it just matches away…). What’s even worse is that Shiri et al. (••) figured out a way to actually restore the original image from the blurred one. So now, you can not only figure out who the blurred person is, you can even get back to the original image from this. Ouch!
A Way Out?
When you get down to it, there are two possible ways out — blackboxing, and inpainting
When you get down to it, there are two possible ways out — blackboxing, and inpainting

InpaintingHere, you swap out the face with a different face, but one that works in context (Technically speaking, inpainting refers to the process of replacing lost/corrupted parts of an original image/video. Done well, this can be awesome. Done poorly, well, you can end up with what happened to Elias Garcia Martinez’s ‘Ecce Homo’.
The key to inpainting for anonymization though, is that you use it to swap in a different face! Mind you, to prevent it from looking goofy, you want to try and match certain things, like face orientation, body pose, activity, lighting, and so forth.
It’s all a lot harder than you’d imagine, and has, hitherto, largely been a manual thing (if done at all).
It’s all a lot harder than you’d imagine, and has, hitherto, largely been a manual thing (if done at all).
A new paper by Sun et al. “Natural and Effective Obfuscation by Head Inpainting” manages to do exactly this though! The best part is that they managed to come up with a target-generic solution (remember blackboxing above? Works against all identification mechanisms?), that still seems extremely natural.
Deep Learning FTW! Now, to move on to video, right?
Comments