The Facial (Un)Recognition Filter

A rather surprising number of people from various social media platforms as well as police surveillance systems are highly interested in what you look like.

 

What this mean for you, is that every time your face pops us, companies are capable of learning a whole lot about you, such as how you look, how old you are, your ethnicity, and much, much more.

 

Although generally this data is used for the purpose of targeting advertisements,  it can also seriously people people at risk to have their identity stolen or privacy violated.

 

In order to help people keep whatever privacy they may have left, there are certain tools that came up out that are designed to prevent facial recognition, such as Face Off Hats, which is a hat with a trippy pattern, or 3D masks that confuse face-scanning software. They trick the camera with optical illusions.

 

These are the tactics that throw off the cameras themselves, but if a photo of you ends up online, then that’s a different story, but it’s also got a smart solution in the making.

 

Soon, a new filter may be released that will prevent AI from recognizing your face.  The filter was built by engineers from the University of Toronto that alters people’s faces in photo, to prevent the software to realize what it’s seeing.

 

The AI filter searches for certain facial features, and changes pixels. People can’t really notice the difference, but the AI that’s scanning the imagine won’t be able to tell it’s even a face, let alone your face.

 

In a test, the filter succeeded to reduce to the accuracy of face-detecting from always being correct to only one in 200.