History of Computers - Facial Recognition

Introduction

Facial recognition was pioneered by Woodrow Bledsoe along with others in the early 1960s; however, it has come a long way since then. With the evolution of 3D imaging and rendering software, facial recognition is much more highly advanced than it was during the 2D era, where someone would have to be looking directly at the camera for the system to work. [1]

Overview

Facial recognition is a non-intrusive, biometric identification method, by which a digital camera captures an image of an individual and analyzes their facial features to determine similar or even perfect matches from a stored database. This system is often used for threat detection, as well as a standard identification method. The facial recognition system is marketed as a highly successful threat detector and identification method, yet it is far from commonplace. However, many project this system to overtake fingerprint identification in the coming years [2], which can be seen in Apple's new iPhone X, where they swapped their fingerprint scanner for a more advanced facial recognition system.

How It Works

Facial recognition works fairly simply through a series of processes to identify the subject. First, an image of the subject is detected through a live video feed or from an uploaded image. Second, the system detects the size, pose, and alignment of the face up to 90 degrees. Third, the system measures the curves and structures of the face on a sub millimeter level to create a template. During this process, over 80 nodal points are measured: including distances between eyes, nose, mouth, jaw edges, cheek bones, and other facial features. Next, the template made in the previous step is translated into a unique code, in which the numbers represent the subjects various facial features. Fifth, the new unique code, which comes from a 3D image, is run through a 3D facial recognition database to find possible matches. However, if the database is an outdated 2D database, the code must be converted/altered so that it can work with the old database. Finally, the image is matched to only one database image, which obviously represents the inputted image the best.[3] Facial recognition is also difficult to fool as it measures facial landmarks, meaning it's difficult for you to cover up with glasses, a beard, or makeup. However, a full on mask could still do the trick. [4] Modern facial recognition are developed by supervised Deep Learning networks. Deep Learning networks are a type of AI.

Significance

The significance of facial recognition is that it has prevented many threats worldwide, and has the potential to prevent many many more. In the near future it may take over as the most prevalent biometric scanning device, and we have already seen that through the iPhone X. It is also now being used in government buildings worldwide, such as a local DMV, as well as in banks and other high-security areas. While facial recognition has worked to improve security nationwide, it also serves various other purposes, such as facial recognition to match you to a celebrity, sports player, and more.

References

http://www.ex-sight.com/technology.htm

http://electronics.howstuffworks.com/gadgets/high-tech-gadgets/facial-recognition2.htm

External Links

[5]

[6]