Deep learning and the future of facial recognition

Facial recognition technology is being increasingly used in a range of use cases. But it has its pros and cons. In this infographic, we outline some of the ways facial recognition software is being used and why.

Infographic transcript

What is deep learning?

Deep learning enables machines to learn and solve complex problems using algorithms inspired by the human brain without any human intervention.

Deep learning algorithms need data to learn, and lots of it! But that’s no problem because we generate approximately 2.6 quintillion bytes a day1.

Facial recognition

Facial recognition uses images captured of an individual’s face from photos or videos.

The distances between the eyes, nose, mouth and jaw are measured.

Those distances are converted into a digital algorithm.

The algorithm is compared to a database of previously analyzed faces until a match is made.

Facial recognition is big and the global market is expected to rise to USD 9.6 billion by 20222.

Facial recognition and security

Russia, India, USA, UK and China are all developing, trialling and using real time facial recognition3.

Some Chinese security forces wear glasses with facial recognition technology built in which they can check against their database4.

In Wales, facial recognition technology helped the police identify and arrest over 450 individuals between the summers of 2017 and 20185.

Facial recognition in the fight against cybercrime

Facial recognition can help to positively verify and authenticate people, mitigating against the current cyber crime crisis.

The global cost of cybercrime in 2018 rose to $600 billion — about 0.8% of global GDP6 and it is estimated to rise to $6 trillion by 20217.

Passwords alone are not secure. 81% of hacking related breaches leveraged either stolen and/or weak passwords8. Facial recognition provides an added layer of security

Facial recognition technology is difficult to hack but it is not immune. Hackers won’t stop in their quest to replicate faces or hack into devices and systems that store data relating to our faces. That is why two-factor security is best.

Facial recognition for convenience

Facial recognition technology for payment is in its infancy but has growing support.

35% of Americans9 and almost 25% of Brits10 support facial recognition payment technologies.

Cool uses for facial recognition technology

Customers at KFC in China can simply smile at a screen to pay using facial recognition11.

BBVA bank is working on linking RFID (radio frequency identification) tags in products into the payment system, so customers can choose items in-store, look at the camera and pay12.

Professor Shen Hao of the Communications University of China uses facial recognition technology to track students’ attendance13.

ESG Management School in Paris is using facial recognition software in its online classes to analyse eye movement and make sure students are paying attention14.

Tesco plans to install screens in petrol stations across the UK that will scan customers’ faces to determine their gender and age so it can run tailored ads15.

Helping Faceless16 is an app using facial recognition to find missing and kidnapped children.

Finding Rover17 is an app using facial recognition to find missing pets.
China Southern Airlines18 and British Airways19 are allowing passengers to board aircrafts by scanning their faces.

Supermarkets in the UK will soon be using facial recognition systems to determine age and approve sales of alcohol and cigarettes20.

Potential pitfalls

Racial bias

Facial recognition tools are much more accurate in identifying light-skinned men in comparison to darker-skinned women21.

It can be fooled

Obscuring the face with hats and scarves makes it difficult for facial recognition technology to get an accurate reading.

Artist Adam Harvey has developed state of the art hairstyles, cosmetics and scarves to fool facial recognition algorithms22.

Civil liberties

Few laws regulate facial recognition and surveillance. Campaigners argue it is an infringement on civil liberties and the rights of ethnic minorities who are more likely to be falsely identified23.