New Technology and it’s Impacts and Improvements
on an Organisations Operations
Developments in Facial (and Biometric) Recognition (P5)
Types of facial recognition software and hardware have been around for many years but has recently
become increasingly popular with organisations, as well as throughout the public due to apps that use
facial recognition. Earlier versions of facial recognition were a lot simpler, they function via a camera
viewing someone’s face and using the measurements such as the distance between a person’s eyes, and
general facial proportions – then using an algorithm to attempt to match this to pictures of people in a
database. However, this was not always effective as the main use (and one of the only uses) for facial
recognition several years ago, was for security cameras which are commonly low resolution and do not
have to access to a lot of computer power with a strong, complex database of people. Databases of people
were usually restricted to people with criminal convictions or who are on a wanted list, with 1-3 still
pictures, this means the software was not useful for many applications and the low-quality meant
identification was easily avoidable and often unsuccessful due to surveillance systems not being able to
detect differences made by facial expressions, multiple angles, movement, and difficult lighting. An
example of how poor facial recognition was, is that according to the BBC “police were only able to identify
one person out of 4,000 images that were taken during the London riots in 2011”.
One of the biggest developments in facial recognition is that instead of
using 2D images, software has been improved to use 3D models, this
allows for countless angles of users and has fewer restrictions caused
by lighting issues. These models/images can be converted between 3D
and 2D at any time without losing data, the type of biometric data this
new software can detect include bone structure, curvature, and facial
expressions. The example of facial recognition software I will be using
is called Face++ (face plus plus); it was developed in China, Beijing by a new start-up who are now valued
around one billion dollars with plenty of investment from organisations and the government. The database
of users holds hundreds of millions of users, with 3d models of their face with movement and facial
expressions detectable, as the software can use a ‘liveness test’ to prevent duping the system by a user
having to move and/or speak while scanned. This type of advanced identification has countless uses from
providing authorization for any payments, providing access to facilities, and of course, to detect criminals in
public settings. Face++ has built up its massive database of users by being used in popular apps, like apps to
transfer money such as Alipay with 120 million users in China, that uses just your face as recognition. It is
also used for Didi, China’s version of Uber, where users can use it on the drivers to ensure they are
legitimate. Facial recognition can be used confidently by organisations and the government due to highly
accurate, enough so payments and security can depend on
it with software like Face++. DeepFace is another system
created by Facebook, which also uses 3D face modelling
and complex layered software to achieve an accuracy of
97.35%, and is used for tagging people in photos on social
media, with a huge database trained on four million
images belonging to 4,000 people.
on an Organisations Operations
Developments in Facial (and Biometric) Recognition (P5)
Types of facial recognition software and hardware have been around for many years but has recently
become increasingly popular with organisations, as well as throughout the public due to apps that use
facial recognition. Earlier versions of facial recognition were a lot simpler, they function via a camera
viewing someone’s face and using the measurements such as the distance between a person’s eyes, and
general facial proportions – then using an algorithm to attempt to match this to pictures of people in a
database. However, this was not always effective as the main use (and one of the only uses) for facial
recognition several years ago, was for security cameras which are commonly low resolution and do not
have to access to a lot of computer power with a strong, complex database of people. Databases of people
were usually restricted to people with criminal convictions or who are on a wanted list, with 1-3 still
pictures, this means the software was not useful for many applications and the low-quality meant
identification was easily avoidable and often unsuccessful due to surveillance systems not being able to
detect differences made by facial expressions, multiple angles, movement, and difficult lighting. An
example of how poor facial recognition was, is that according to the BBC “police were only able to identify
one person out of 4,000 images that were taken during the London riots in 2011”.
One of the biggest developments in facial recognition is that instead of
using 2D images, software has been improved to use 3D models, this
allows for countless angles of users and has fewer restrictions caused
by lighting issues. These models/images can be converted between 3D
and 2D at any time without losing data, the type of biometric data this
new software can detect include bone structure, curvature, and facial
expressions. The example of facial recognition software I will be using
is called Face++ (face plus plus); it was developed in China, Beijing by a new start-up who are now valued
around one billion dollars with plenty of investment from organisations and the government. The database
of users holds hundreds of millions of users, with 3d models of their face with movement and facial
expressions detectable, as the software can use a ‘liveness test’ to prevent duping the system by a user
having to move and/or speak while scanned. This type of advanced identification has countless uses from
providing authorization for any payments, providing access to facilities, and of course, to detect criminals in
public settings. Face++ has built up its massive database of users by being used in popular apps, like apps to
transfer money such as Alipay with 120 million users in China, that uses just your face as recognition. It is
also used for Didi, China’s version of Uber, where users can use it on the drivers to ensure they are
legitimate. Facial recognition can be used confidently by organisations and the government due to highly
accurate, enough so payments and security can depend on
it with software like Face++. DeepFace is another system
created by Facebook, which also uses 3D face modelling
and complex layered software to achieve an accuracy of
97.35%, and is used for tagging people in photos on social
media, with a huge database trained on four million
images belonging to 4,000 people.
