top of page

A brief guide to facial recognition technology in Asia

by Anh H. Nguyen

Some key points:

-Facial recognition is the ultimate in identification technology.

-Asia may pull ahead of the US and EU countries in the facial recognition race

-Barring applications in policing, facial recognition could support retailers in unexpected ways.

Facial recognition is the face of the future, and for good reasons

Which is the most natural kind of biometrics? Facial, of course. Most of the time people recognize one another by looking at faces first. That is why other bodily signatures such as voices, postures, fingerprints, irises, and DNA take a back seat to facial features when it comes to identification tech. Not only that, facial measurements are usually the most simple to deploy and implement, requiring no active interaction from the end users - most of the time, they are not even aware that their facial characteristics are being detected, captured, and analyzed. Moreover, extensive camera systems installed in public places around the world certainly make it easier for facial recognition softwares to do their job, since they provide ample amounts of samples for machine learning and come in handy as pre-existing hardwares. And the processes for verification are usually more expedient than, say, DNA tests, for instance: with a database large enough, face matching could be instantaneous.

Being the number one in the biometrics sector, facial recognition technologies live up to their reputation by improving all the time. Theoretically, AI technologies have already surpassed humans when it comes to face ID since 2014: GaussianFace, an algorithm developed by the Chinese University of Hong Kong scored a 98.52% performance compared to a 97.53% achieved by people. Accuracy has also risen at an exponential rate: in this NIST report, failure rate to match among 26.6 million photos decreased 20 times over a period of 5 years from 2013 to 2018. For an example closer to home, Google's FaceNet has a stunning 99.63% rate of correct face matching.

The Eastern and Western worlds hold diametrically different attitudes when it comes to facial recognition

This study published in 2019 estimates the facial recognition market to reach $7.0 billion valuation by 2024, but the actual figure could be much higher. Because of its huge potential, the field of facial recognition is exceedingly competitive with new theories and innovations being developed all the time. All the tech giants in the US, Google, Apple, Facebook, Amazon, IBM, and Microsoft pursue R&D in facial technology to varying degrees - quite unsurprisingly, given their big pockets and diverse interests. North America also holds the largest market share during this period. However, the top vendors in the facial recognition market represent quite a healthy mixture of countries, including NEC (Japan), Aware (US), Gemalto (Netherlands), Idemia (France), Cognitec (Germany), nVviso SA (Switzerland), Techno Brain (Kenya), Neurotechnology (Lithuania), Innovatrics (Slovakia), Herta Security (Spain), Megvii (China), and SmilePass (UK). This shows that developed countries do not have a monopoly on making headway in biometrics tech and Vietnam may very well hope to harbor some world-class AI startups in the future.

When people think about AI facial recognition, surveillance often comes to mind first. Since the '60s days of Star Trek, Hollywood sci-fi movies have long thrilled viewers with ominous, albeit exciting glimpses of futuristic surveillance states where people could neither run nor hide from all-knowing, all-seeing face scanners. Alas, it is Asian countries that have managed to turn those visions into reality. Right now, China has the most widespread facial recognition system in the world. It could only be described as both omnipresent and omnipotent, logging almost every single Chinese citizen, even ones as young as 9 days old, which is no mean feat considering China is also the most populous country on earth with nearly 1.5 billion people. A leak from the database of a surveillance company in China called SenseNets showed it recorded 6.8 million locations in one day, just by using real-time facial recognition to track people. And that is just one company.

Ongoing debates about the scope of data privacy related to facial recognition have been a hot topic since day one. People are justified in worrying about the intrusions and violations of their private lives, with fear of authoritarian governments looming overhead. Laws and regulations regarding the use of facial recognition are all over the place, with outright bans in some countries and policy vacuum in others. Even in the same country, legislations could vary widely. In the US, for example, there are only 5 states that formally protect the biometric data of their citizens. The remaining 45 states still function in a gray area. To complicate matters, some cities in the US opt to forbid facial recognition completely, including Boston, San Francisco, Portland, and San Diego. In the UK and Europe, rules governing data protection of citizens are even more rigorous.

Clearly, facial recognition has an interesting position: it is embraced by the legal systems in some places, and repelled by others. This uneven patchwork has far reaching consequences on the progress of facial recognition technology. Three of US Big Tech companies, IBM, Microsoft, and Amazon have either temporarily suspended or stopped both their R&D activities and sales of facial recognition solutions to the police and other law enforcement agencies. In light of the heated political climate in the US, this could be a knee-jerk reaction intended to soothe the community. Whatever the motivations may be over there, not being restrained by laws and public sentiments gives tech companies in other regions a considerable advantage. Perhaps not surprisingly, although North America still leads the facial recognition market in terms of share, Asia holds the first position in terms of growth.

A few salient facts:

-There are more than half a billion CCTV currently in use in China. Of the world's top 10 cities with the biggest number of street cameras per person, the first five are all China's cities.

-The Tokyo Olympics 2021 will use facial recognition to improve security by identifying athletes, staff, officials, and the media. The system is being developed by a Japanese tech firm named NEC Corps.

-The Aadhaar program in India is the world's biggest biometric database with details about 1.26 billion individuals. So far, Aadhaar has saved the Indian government $12.4 billion in operation costs.

-In other countries like Brazil, Russia, and some African nations, projects to collect biometric data are also underway.

-CCTV cameras are a part of Vietnam's ambitious plan to build smart cities. Ho Chi Minh city especially has widely adopted an AI-based surveillance system to help with managing traffic.

Retail is an overlooked beneficiary of facial recognition technology

Feeling a bit wary about the loaded implication of AI surveillance? Fret not! Since many of the discussions around facial recognition are dominated with concerns about data protection and privacy, people tend to let surveillance overshadow other sectors that it could serve. Retail is the area that could possibly gain the most from the use of facial recognition, yet it has escaped the spotlight for quite a while. Here is an incomplete list of the many ways facial recognition could change retail:

-Many shoplifters and other unscrupulous characters are repeat offenders. When retailers use facial recognition to help preventing thefts and other misdemeanors, once people violate the stores' policies, their details could be detected and documented. In case they return to the stores, the camera surveillance system would recognize them and swiftly notify the security personnel. With a facial recognition system in place, retailers could rest assured that they are doing their best to minimize losses and ensure a safe environment for both their employees and customers. FindFace, for example, touts an "up to 75%" of petty theft losses reduction with the shoplifter protection it provides.

-With facial recognition solutions, retailers could get the most detailed & precise report of unique visitors, new customers, and repeat customers. This could help them assess the effectiveness of their marketing campaigns, sales tactics, visual merchandising, and loyalty programs.

-Retailers could get to know their customers at an unprecedented level. When used in conjunction with customer analytics tools such as Palexy's Store Optimizer and POS data, facial recognition softwares could yield a thorough picture of customers' experience, behaviors, preferences, and shopping history. From there, retailers could keep improving all manners of operations to expand their customer base and meet other KPIs.

The "virtual sale assistant" (VSA) empowered by facial recognition could transform retail as we know it

To expand on the two previous ideas, retailers could identify VIP customers immediately after they walk in and tailor their services to meet their highest expectations. This has the double benefit of both increasing sales and customer experience and will be particularly useful for high-end retailers. In countries that do not inhibit or limit the use of identification tech, these retailers would find it worthwhile to explore the many facets of VSA, which could range from basic to sophisticated. In this article, we further highlight the power of such a VSA for the luxury sector.

As competition in the retail market becomes more fierce, though, personalized shopping will become more mainstream and not reserved exclusively for premium brands. Who is to say only fancy establishments could provide good, customer-centric services? In the era of mass-produced goods and online shopping, brick-and-mortar retailers usually struggle to attract new customers and, more importantly, keep drawing them back. Even mid-market and small businesses will find that enhancing customer experience not only helps them stand out but also sustain frequent visits better than any marketing campaigns.

That possibility did not become feasible until the arrival of facial recognition due to costs and other logistics issues. For the first time now, shopping personalization has become scalable. In the future, this may even become a necessity as more and more retailers adopt AI tech to provide the best services possible.

What is in store for Vietnam's retailers?

After an erratic, sometimes depressing year, the upcoming chapter seems to contain a more positive note for businesses as people's lives return to normal. Christmas and New Year's Eve are right around the corner, while the Lunar New Year, Valentine's Day, and March 8th will follow in quick succession - the golden period for shopping. Now is the ideal time for retailers to try out new solutions to boost their year-end sales and set new goals for 2021. Powered by Palexy's trademark facial recognition technology, the Store Wizard is the latest in our product lineup. Like its predecessor the Store Optimizer, the Store Wizard is designed to help retailers target and optimize all areas of business. Contact us today at



bottom of page