Saturday 31 August 2013

Face Recognition Technology Advancements

The use of Face Recognition Technology has been traditionally dominated by the security industry. Video surveillance and identification applications are widely used in airports, and in police and army facilities.

Nevertheless, a new civic branch of face recognition is emerging. From demographic analysis applications such as counting people in the crowd, telling their sex and age and even mood, through social apps, games and even a smartphone app that recognizes celebrities, face recognition revolutionizes old ways of doing things.

Overwhelming advancements in technology over the years have allowed the development of new applications in almost every field. Following are the major advancements in technology that are also most relevant to face recognition:

 

From Commodore to 6 Core

 

With computing power doubling itself almost every two years for the past decades (Moore’s law), we have come a long way since the PC Commodore 64 had hit the market 30 years ago, with its 1MHz processor and 64 Kilobytes of RAM. A rough comparison to Intel’s i7 6 core would show that computing power was multiplied by 20,000(!) since, and the increase in RAM skyrocketed even higher. In addition (and that’s a big addition) cloud computing and multi-threading provide scalable and relatively cheap solutions for high volume real time systems.

 

More Pixels Less Dollars


The continuous sharp rise in the “pixels per dollar” ratio allowed the proliferation of megapixel cameras which serve as the “eyes” of a face recognition technology.

 

Smartphone Anyone?


What was once considered to be a super-computer, now literally falls ages behind your average smartphone. Armed with a fierce processor, hi-res screen and camera and an internet connection, the smartphone is everywhere and it changes the way human interact with each other and with their environment, allowing the development of new apps that weren’t possible thus far.

 

Improved Accuracy


In addition to the increased computing power, face recognition algorithms accuracy has improved substantially in the last 20 years (from 80% to 99% in a control environment according to NIST), making it much more reliable and robust.

 

Face Recognition API


An API (application programming interface) is usually a simple way for a software product to connect and run an external software component. The availability of face recognition API’s allows developers to focus on developing new applications and solutions by utilizing an existing face recognition technology rather than developing the face recognition technology by themselves.

No comments:

Post a Comment