From security to toilet paper, China has played the facial recognition to the hilt

It is estimated that the global facial recognition technology is a USD178 million business in 2016, though accounts for only 4% of the total biometric market cake while fingerprint still dominates 91% of the revenue. But China is going to ride this wave and build its own momentum&scale, very likely surpassing any western countries.

Conventional wisdom considers facial recognition falls into 3 basic categories along with few specific applications.

China in fact already adopted facial recognition in most of the areas. Let us just have a brief rundown of those deployments which caught the attention of news headline:

  • KFC allows customers pay for orders with their face powered by Alibaba. It is said the technology also suggests food choices based on customer’s age, sex, and mood
  • The sudden hype of unmanned convenience stores in China is leveraging facial recognition capability
  • China Southern Airlines recently ditched boarding pass by introducing facial recognition in certain city; Chinese railway stations are testing a similar system to replace train tickets
  • Many local banks gradually rolled out facial recognition ATMs across cities
  • A least four cities in China have installed facial recognition technology to name and shame jaywalkers
  • Police used the technology to arrest drug abusers during a beer festival event
  • In Shenzhen, the city government tested the device in 12 taxis that verifies the ID of a driver
  • Beijing has integrated facial recognition into the public toilets in a park to identify “toilet paper thieves”

 

 

Below video should give you a vivid demo of different kinds of facial recognition application in China, including that “toilet paper dispenser”.

 

 

Facial recognition mechanism basically

  1. begins with finding a face in an input picture or video: an algorithm searches the input picture for groups of pixels arranged in patterns indicative of a human face.
  2. It next moves on to feature extraction: it relies on a large training set. So whenever a new picture is thrown at the algorithm, it calculates the distance between the newcomer and each of the pictures it has been trained on. It ends up with a feature vector, a numerical value corresponding to the way we look. The latest deep learning algorithm obviously makes the whole training process more efficient and faster.

 

WSJ also explained this trend so well in this video.

 

 

 

China offers the fertile and unique environment to cultivate the advancement of facial recognition:

I)One key factor driving the growth of the facial recognition market is the investment in surveillance systems. For that, China has 170 million surveillance cameras in public places and is likely to install an additional 450 million cameras by 2020.

 

II)China government has the centralized facial database for all the citizens over 18 years old and is very willing to partner with internet giants to implement the technology which adds god-like power to assist its homeland security and criminal investigation.

 

III)In the West, the public often raises serious concerns about privacy and misuse of personal information. For instance,

  • In 2016 in the US over 50 civil liberties groups protested to the Justice Department’s Civil Rights Division regarding the growing use of facial recognition technology by police forces.
  • In 2016 Snapchat was sued for illegally collecting, storing, and using biometric information, including face geometry and iris scans of its users without obtaining permission.
  • When Walmart filed a patent application to develop facial recognition at the checkout to help store personnel determine if customers are happy, critics said this should be the worst AI application ever by violating personal privacy.

These are NOT problems in China, folks!

 

IV)Tons of money from the local capital market have poured in. Just look at two unicorns in this area:

  • Face++, founded in 2011, the technology provider behind Alibaba’s “Smile and Pay” for KFC raised USD100 million in Series C.
  • Sensetime, a 3 years old startup with core technology in the face, has received USD410 million in Series B.

 

Infrastructure, data, willingness, money, bingo…All the stars are aligned to make the accelerated adoption of facial technology in as many ways as imaginable in China.

 

Author: Cecilia Wu