AI Can Recognize Your Face Even If You’re Pixelated

Pixelating and blurring images has long been known to be insufficient in completely obscuring information [1]. In fact, in a lot of computer vision work, the image resolution is reduced to ignore noise and facilitate the workload for algorithms anyway. Completely destroying information through blacking out is preferable.<p>[1] <a href="https:&#x2F;&#x2F;dheera.net&#x2F;projects&#x2F;blur" rel="nofollow">https:&#x2F;&#x2F;dheera.net&#x2F;projects&#x2F;blur</a>

I’m not terribly surprised by this. Downsampling is a really effective way to create a “fingerprint” by which you can identify some data. And that’s essentially how an image-searching service works.

I dont know about pixelation, but gaussian blur can be undone mathematically. Pixelation at least destroys the vast majority of the information.

if you have the original font, blurring can be defeated . always better to just black out text that you don;t want to to be seen

Can journalists stop writing &#x27;AI&#x27; everywhere when it&#x27;s just Neural Nets? It&#x27;s all starting to look ridiculous - if you need a popular science friendly word, what&#x27;s wrong with Image Recognition Program?

Reminds me of <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Christopher_Paul_Neil" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Christopher_Paul_Neil</a>

This is based on having the original (unpixelated) image in every case, right?