The Lowdown On China’s Facial Recognition Technology

The march of technology, it seems, cannot be denied.

2019 has been an exciting year – data analytics, artificial intelligence and machine learning have become increasingly accepted as the defining buzzwords of our day as businesses become increasingly willing to invest in technology in the name of increasing their returns on investment.

Elsewhere, Augmented Reality (AR), Virtual Reality (VR) and Mixed Reality (MR) continue to play important roles in industries where small errors have huge costs. Think space exploration, engineering and even healthcare. But more interestingly, as technology has become more affordable, AR, VR & MR have moved away from their specialised uses – and landed squarely in the public eye. Marketers rejoice as AR, VR & MR enter the fray in a never-ending war for brand equity – which means we can prepare ourselves for bigger, better and more exciting advertising experiences in the near future.

But if we’re talking about technologies that’ve moved away from their original uses and into the public eye, then let us consider the humble face filter.

Face filters are but a fairly contemporary subset of a facial recognition research movement originating in the 1960s. Originally motivated by security concerns, it has since made a successful transition into the public eye, and facial recognition has now been put to much loftier and nobler ends. We’ll talk more about the beginnings of facial recognition later.

Pictured: Loftier and Nobler ends.

But this frivolity is powered by literally billions of machine-learning hours – hours spent identifying and mapping the human face thanks to an entire civilisation of willing social media-enabled guinea pigs. And while one might scorn the budget-store Dr Moreau who posts pictures of themselves with outsize canine ears on Facebook, one should also consider the many uses of this facial recognition technology in everyday life – some benign, others significantly less so.

And of course, no discussion of benign and un-invasive technology use is complete without everyone’s favourite global powerX with a heart of gold.

Now let’s see what China has been up to.

Big dreams await China

Based on this graph here, Sensetime, a Chinese security company, is dominating the startup landscape. According to Forbes, Sensetime’s core business comprises face, image, object and text recognition, medical image analysis, remote sensing and autonomous driving systems. Their solutions are widely used across industries as diverse as healthcare, finance, security and retail.

Sensetime might be seen as a microcosm of China’s ambition – an AI world leader come 2030. And if we’re reading the signs right, this is no empty boast.

In the field of facial recognition alone, China launched an ambitious project in 2017 – to create a system capable of identifying any one of their 1.3 billion citizens immediately within 3 seconds. Right now, nothing of this scale exists in China but according to South China Morning Post, security cameras with facial recognition capabilities were deployed in Tiananmen Square as early as 2003. We reckon that’s something that any halfway paternalistic regime in the world would trade their two front teeth for.

In sum, China is doing fantastically well where facial recognition is concerned and even has intentions to triple its takings to $300 million in the following year by turning SenseTime into a national champion.

But it’s not just the government that’s been pushing facial recognition technology aggressively. Private companies and merchants in the nation also play a prominent role.

Applications of China’s facial recognition tech

Forget NETs, PayNow and payWave – if we’re talking cashless – China is the real deal.

Be it shopping, transport or medical bills – the Chinese pay with their faces.

Just two years ago, MIT Technology Review lauded face payments as one of the top 10 breakthrough technologies globally. While this does raise a series of disturbing questions pertaining to impersonation and calls to mind a certain 1997 action film starring John Travolta and Nicholas Cage, we cannot deny the convenience that must surely come with pouting into a camera to pay for a bottle of Fiji Water. Bravo, China!

But don’t take our word for it – here, have a screenshot:

But it gets better. What could be worse than pooping in a public toilet without ascertaining the status of the toilet paper dispenser? Better people than I have had to crab-walk gingerly out of the cubicle (come on, you know exactly what I’m talking about) in search of processed fibres with which to wipe.

This is not a problem in China, though.

In China, toilet paper dispensers are smart enough to remember someone’s face and set a limit on the number of squares of TP you get in an hour.

To some this might be overkill but if you’ve ever had your blood run cold at the end of a rewarding series of bowel movements you’ll understand that this is completely justified.

Let’s just hope citizens don’t start a ruckus because they have to fold something in half more than 8 times.

The technology has also been implemented in subway stations and at school gates.

Starting April 1stthis year, passengers in Jinan in Shandong province can enter subway stations with a smile. Literally. Jinan Metro has done away with the need for tickets or cards. Passengers need only smile at the camera to gain entry.

Meanwhile, schools are also tightening their security by enforcing the technology at school gates. At the same time, it has also been used to monitor students – from stopping fights to making sure they pay attention in class.

Along a similar vein, a handful of police departments are using facial recognition to analyse video footage and identify criminals.

Even casinos are joining in the fun and using the technology to their advantage. Catching hardcore gamblers and identifying cheaters? Check and check. Facial recognition has got all these casinos’ back.

We just really hope no one puts all this data to nefarious purposes, of course.

But as eager as China is to assume the mantle of a world leader, her ascension will not go unchecked.

The US is stepping in again

The intense rivalry between the US and China has finally boiled over. The US-China trade deficit has long been a source of tension between the two countries. And things only got worse when the Trump administration accused China of stealing commercial technology and intellectual property earlier this year

The US’s answer to China’s ‘crime’ was the Huawei ban, with Google following up with a suspension of Android OS updates and certain Google services. The world went nuts when the news broke – particularly Huawei themselves and Huawei users. After all, who can bear to say goodbye to Google Play, Maps and Gmail?

But China’s not going out without a fight.

The word out on the street is that Huawei is currently developing its their own OS, Hongmeng, or more aptly, ‘Harmony’, claiming that it’ll be up to 60 percent faster than a typical Android phone.

Even then, the ban has hit Huawei hard, with smartphone sales dropping an estimated 40% to 60% and revenue falling by about $30billion.

The US’s plan to dominate China in the tech industry continues with Hawaiian Senator, Brian Schatz proposing the “End Support of Digital Authoritarianism Act” to forbid companies from China, Iran, Russia, North Korea and other countries that always violate “internationally recognised human rights” from the Face Recognition Vendor Test (FRVT), which is known as the gold standard for establishing the reliability of facial recognition technology.

Facial Recognition: Origins

Let’s go all the way back to the 1900s, where facial recognition tech was still in its infancy.

Some (including but not limited to Wikipedia) would say that American mathematician Woodrow Wilson Bledsoe was the father of facial recognition. Some would be right.

In the 1960s, he developed a RAND tablet, a system which was able to input vertical and horizontal coordinates on a grid with a stylus which released electromagnetic pulses. In other words, it was also able to manually record the specific coordinate locations of different facial features like the hairline, nose, eyes and mouth.

When this technology was made public, it created quite a buzz. But it was early days yet.

In the 1970s, Goldstein, Harmon and Lesk stepped forward to add increased accuracy to the system – specifically with 21 additional markers. Even then, the technology still needed some work.

It was only around the 1990s that the Eigenface approach came into the picture – the development of a set of mathematically averaged faces by the name of Eigenfaces through the application of linear algebra. Simply put, they ran a couple of calculations on a bunch of faces, identified where facial features were located mathematically and averaged everything out to create a genetic face which they would then modify by varying the variables by a percentage.

The rest, as they say, is history.

Enter the Future

Now that facial recognition technology is market-ready; governments are also installing the systems in airports to keep an eye on cross-border movement.

For instance, the US Department of Homeland Security has a system advanced enough to identify travellers under criminal investigation or who have overstayed their visas.

One of the most famous uses for facial recognition is, of course, Apple’s iPhone. We’re pretty sure you’ve heard of Face ID – it will scan the shape of your face using infrared light to come up with a depth map of your face. If it matches whatever they have on file, your phone unlocks.

No effort needed.

Law enforcement and military professionals are also using face recognition. It’s particularly useful for identifying corpses. Fun fact: remember Osama bin Laden? It was thanks to facial recognition that authorities were able to identify him after he was killed in an American raid.

Some have argued that the human maintains its position at the top of the food chain by virtue of innovation. That’s what happens when you take two separate items, mix them together, and end up with another item. For instance, tying a rock to a stick to make an axe. For a more contemporary example, consider how drones have been married with facial recognition technology to create an even more invasive system of surveillance capable of following you around. A 10kg lens mounted on a drone allows the system to spot a suspect from approximately 800 metres away.

So there you have it – China’s face tech, in short. But the interesting thing about technology is that improvements in one area tend to precipitate improvements in other spaces. I’m looking at you, deep fakes. And we’re willing to bet that the face of the animation and motion graphics industry will evolve even further in the very near future.

So, China Inc’s definitely growing. We don’t need an animated explainer or an infographic video to tell us that.

What about you? What’s your next move as a business?

Leave a comment

Your email address will not be published. Required fields are marked *