Technology Top News

Surveillance evolves as cameras get a brain

Advanced new facial recognition technology is giving ordinary surveillance systems the ability to learn, analyse what cameras see, conduct market research and even proactively mitigate risk.

Video Surveillance systems are an important step towards enhanced security, but they have certain shortfalls – they are very dependent on the ability of the control centre and monitoring team to be alert to every moment on every screen, all the time. “With surveillance cameras now in place virtually everywhere, it becomes extremely difficult to analyse who is where and what they are doing, all the time,” says Marius Coetzee, CEO of SA-based identity specialists Ideco. “This means traditional surveillance footage is typically used after an event to support investigations.”

However, advanced new facial recognition technology has come to market to add intelligence to surveillance systems, allowing for proactive monitoring and control 24/7, automated facial wrapping and recognition, and even highly accurate identification.

Facial recognition is a fast-growing global market expected to be worth over $10 billion by 2025, as organisations seek improved speed and accuracy in security systems.

“It’s a complex environment, and there is some confusion in the market about what constitutes facial recognition,” says Coetzee. “Essentially, advanced new intelligent facial recognition technology uses big data analytics, AI and machine learning to capture faces, establish factors such as their age, gender and ethnicity, analyse their facial expressions and even match and identify the person. And because it is constantly learning, the more you ‘train’ the technology, the better it becomes. There is massive potential for this type of technology.”

Coetzee says that by adding intelligent facial recognition technology to existing video surveillance systems, retailers could not just count how many customers looked at a particular area of the store, but also determine how many customers came back more than once, conversions, customer demographics, and how they reacted to the products – based on their facial expressions. This allows for highly accurate market research covering not just customer demographics, but also market sentiment. Retailers could also blacklist the faces of known criminals or fraudsters, or receive notifications when certain high value customers enter their stores.

At stadiums and public events as an example, intelligent facial recognition technology can be deployed to identify troublemakers, manage risk or identify and protect VIPs. At casinos, it could be harnessed to identify regulars, VIPs or even people behaving suspiciously.

For safety and security, intelligent facial recognition allows authorities to list people of interest and be notified when they pass a camera; or secure a region of interest by triggering an alert when an unauthorised person enters that area. It can also complement existing biometric access control solutions such as fingerprint turnstiles, allowing the enterprise to identify which employee committed a crime, for example. At international border posts and airports, intelligent facial recognition is rapidly being rolled out to confirm identity at passport control as part of biometric-enabled “Curb to Gate” airport automation programmes, which are expected to generate $1.3 billion in revenue over the next five years.

“With this new technology, there is so much more information added to the video feed,” says Coetzee. “You can ‘wrap’ a face, extract the features and expressions, produce an identikit of the individual, identify the person and recognise them every time they pass a camera, identify who they associate with and track how they move through the organisation.”

“Because this technology is highly sophisticated, it was previously only in use by big-budget law enforcement agencies. But the barriers to entry have dropped and it is now becoming available to enterprises,” he says.

Ideco recently partnered with international firm IWT to bring their XnapBox intelligent facial recognition technology to South Africa, and launched the solutions to keen interest at the recent Association of Certified Fraud Examiners conference in Sandton last week. “We saw significant interest in the technology because it is unique in the country, has a massive range of applications, adds new levels of intelligence to surveillance investments, and simply integrates into existing systems,” says Coetzee.


By Mbali Sibiya