Face to Face:
Building Trust through Accurate Facial Recognition

It may not be something we realise consciously, but the ability to recognise people is one of the most important social skills for us human beings. In some ways it is a skill we excel at, but in other ways… much less so!

Find out how tremendously helpful computers can be in improving the speed, accuracy and capacity of facial recognition in order to assist humans in creating safer, more reliable systems we can trust.

How fast is facial recognition?

With enough computing power, a machine could recognise every person on the planet in 500 milliseconds. At Innovatrics, our algorithms can match a single person from among 8 billion people in under a minute, using just three standard servers. The same task would take a human almost 100 years!

Humans vs. machines: Who does it better?

The human process of recognising faces is biological, and as such it is subject to moods and biases, and has a limited capacity.

Computerised facial recognition works with analytical algorithms and databases. It is sensitive to some changes, but has a virtually limitless capacity.

Generally speaking, while humans are pretty capable of recognising the same people in different conditions, our reliability can be affected by many different factors, plus there are only so many people we are truly able to remember.

Computers have to be programmed and algorithms put in place in order to achieve reliable recognition, but ultimately, computers are considerably less biased and a lot more accurate when it comes to recognising people.

Further Reading on Facial Recognition

Facial recognition has had immense influence on both society and business in recent years. Learn more about how the technology’s fast development has changed the business landscape, as well as human behaviour, in Kai-Fu Lee’s book AI Superpowers.

Why is this important for building trust?

Yes, computers make mistakes, as do people. However, with the ability to efficiently check, monitor and override faulty machine-made decisions, machine-based recognition of people has already reached insurmountably higher levels of accuracy than people will ever manage to.

SmartFace Platform

Scalable Facial Recognition Server Platform Able to Process Multiple Real-Time Video Streams. Detects and tracks faces in parallel video streams from multiple IP cameras without compromising speed and accuracy.


Having accurate, reliable facial recognition where safety is important can significantly reduce risks of fraud, theft, espionage and more. Governments, banks, airports or ride-sharing services are already all using biometrics to reduce risk and improve safety and efficiency.

AUTHOR: Tomáš Beňadik
PHOTOS: Shutterstock