7 Ways to Overcome Performance Anxiety

Public speaking can be a bigger fear than financial ruin, sickness or death. But even if it isn’t your biggest fear, it’s likely you’ve experienced performance anxiety at some point in your life…

Smartphone

独家优惠奖金 100% 高达 1 BTC + 180 免费旋转




The Hidden Science Behind Facial Recognition

Do you really think it is 100% secure?

Some facial recognition technologies distinguish between pictures of a face and an actual face and some don’t. It depends on the technology being used.

Apple’s facial recognition technology (called FaceID) used to unlock phones is quite sophisticated. It works by firstly illuminating your face with IR light (IR = infrared light that is not visible to the naked eye) and then projecting a further 30,000 (!) IR points onto your face to build a super-detailed 3D map of your facial features. So, it’s the 3D map (i.e. 3D structure) that is analyzed when an attempt is made to unlock the phone. Hence, a photo will not fool this technology.

Facial recognition perceives the world through images or videos (which are a series of images essentially).

A picture of a face is such an image, therefore it cannot tell the difference.

There are supporting technologies you can you to try to determine if a video represents a live person or if it’s a video of a photo, but it’s not the business of a Face recognition system normally.

· Detecting variance in face pose, expression, eyes, etc throughout the video.

· Using additional sensors like an Infrared camera to detect heat, or a 3d camera to detect 3d shape.

· Using active sensors, such as the case with Apple FaceID, which has a micro projector that projects 30K small IR points on the recognized face in order to accurately measure the 3d shape of the face.

Each one of those techniques suffers from some limitations, and in general other advances in AI are quickly getting to the point when they can fool a face recognition system by generating a convincing video of a person’s face.

Samsung’s old phone unlocking technology used to analyze only 2D single frame images (i.e. pretty much what a photo is) of a person to unlock phones. This image would be taken by the front camera of the phone when a person tried to unlock it. As you can guess, a photo could easily fool this technology…

Add a comment

Related posts:

The Return of the Foolish Wealth Tax

The Return of the Foolish Wealth Tax The Wealth Advisor Contributor Mon, 02/28/2022–11:37 Image (Daily Breeze) — In California government, bad ideas are like nightmares. They keep coming back, even…

Cough Medicines for a Sick Patients

Cough is a protective reflex, its purpose being expulsion of respiratory secretions of foreign particles from air passages. It occurs due to stimulation of mechano-or chemoreceptors in throat…

How we started unit testing Boozt iOS app

The codebase of the Boozt iOS app is around 7 years old. Like many codebases of similar age, the code was written largely without unit tests and without testability in mind. Over time, the drawbacks…