China Jaywalking Facial Recognition Guide

By Charles Rollet, Published May 27, 2019, 08:37am EDT

News reports touting the PRC's AI prowess often showcase facial recognition cameras being used to automatically catch and fine jaywalkers.

China Jaywalking Facial Recognition Guide

In this report, we explain how the technology works and examine how common/effective the systems actually are, including:

  • Motivation For Systems
  • 'Board of Shame'
  • WeChat Integration
  • Other Criminal Usage
  • Speed/Accuracy/Effectiveness Claims
  • Providers / Manufacturers
  • Current Scale/Expansion
  • Implementation Problems
  • Conclusion

**********

********** ** * ******** problem ** *****, ** much ** **** ******** ******* ** ** “*******-***** crossing *** ****” (中国式过马路), explaining ****:

** *** **** ****** people, *** *** **, and ******* ** ** with *** ******* ******

* ******** ** *** PRC's ******** ** ****** Security **********:

**** ****** **** ******* safety *********, ***-******* *********, lack ** ******* *** law *********** ******, **** of ******* ************* ** traffic **********, *** **** the ********* ** *************** with **********, *** **** take * ***** *** irrational *************

******** *** **** ** a ******* **** “********* feels * *** **** stepping *** ** ***** of * ****** *****,” The **************** ****, ******:

**** **** ****** (*** the **** ****) *** drive ********** *******, **** scarcely * **** *** passing ************** **** ** pedestrians, ********, **** **, indeed, ***** ****.

** ******* ************* ****** **************** ************** ********** *******, ***** cities **** ******* *************** ****** *** ******** ** 2013. *** ****** *********** systems **** ******* ******* ******* ** ** 2017.

What *******: '***** ** *****'

** ********* ** ******* ******************** *** ****** **** Chinese ***** ***** *********** CGTN:

** ******* ******* ** troublesome *************, *** **** is **** ** ****** jaywalkers. *** ** *** get ******, **** **** image **** ********* ** the ****** ****** *** a **** **** **** right ** **** *********.

**** ** *** '****** screen'. ****'* ******* *** of * ********** ****** by ******* ****** *********** firm*************** * ******* ************ in ********. ** ***** "Previous ****** ** **********" on *** *** *** "Intelligent **** *********** ****** for ******** **********" ** the ******:

********** *************** ******** *** **** Intellifusion ****** ** ********, jokingly ******* ** * 'board ** *****':

Screen Shot 2019-05-13 at 7 14 09 AM

************* ** *** *** only ****** ** **** field. **** **** ********* * ********* '***** of *****' ** *******:

See the signal face recognition system from the "Dong Mingzhu Incident"

** *****:

  • ** *** ****: "****** traffic *****"
  • ** *** *****: "***** the **** ** * civilized ******"
  • ** *** ******: "********** 2018-10-30 **:**:**"

WeChat ***********

** **************, **** ****** **** the "******** ***** ** creating * *** ** peoples' ***** ***** ** photos *** ****** ****'** shared ** ******". **** software ** **** **** to ***** *** ***** of **********.

***************** ****** * ********* living ** ******** ******* that ** ********* *** was ***** ******** ** his ****** ******* ** about "** *******":

** ** ********* ***** *****'* *** ** Facial *********** *****, ******** ** ******* have ******** ******* ** WeChat *** ****** ** use ****** *********** *** payments. ****** ******** * government ** *** ***** number ** **** **, allowing ****** ** ******** link ****** ******** ** individual *******.

WeChat ***** *** ****

* ********** ********* * ******** ****** at ** ************* ** Shenzhen ***** ** ******* of ******, ****** ******* that *** ***** ** jaywalkers *** **** *** crosscheck ** ****** *** then ******* **** ** go *** * **** at *** ****** *******. China *** * ********* system ********** ** *****,** ***** ******* ** photos *** ********* ** authorities ** ****.

* ********** **** ******** ** ********** ***** ***** * text ******* ******** ** a ********* ****** ** one ****** ** ********; the **** ***** ** mention ** ****** *** does ** ****** ****** to ****** ********. ** reads:

[******** ******] ******** ********: You (** **. ********************) have ******** *** *** by *** ********* ******* lights **** ******** *** crosswalk **** ** *** southern **** ** *** East ******* **** ** this **** ** **:** on ***** ** **** and **** ******** *** captured ** *** ******* technician ********** *******. **** record ****’* **** ******* at *** ******. ****** download *** ******** ************** Police *** ** ***** the *********** ******** ** pay *** ****.

No ***** ******** *** *****

** *** *************, ** found ** ****** **** these ********** ******* **** being **** *** ******** but **********. *** *******, while ***** *** ******* concerns ***** *** ********* of **********, ****** **********, etc., *** *** *** of ****** *********** ** identify ****, ** ******* indicate *** ******* ***** used *** **** ********.

** ********* ** * risk, ** *** ******, if *** ********** ******* to ****** **** ** such.

Further ********* *******

***** ********* ** *** faces *** ******** *** not ******** *********, ********** ***** ***** ****** reported**** *** **** * system ** *** **** of*****, *** ****** ****** by ************* ********* ** seconds ** ***** *** taking **** ******** ***** time *** *** ***** is **.

Providers / ************

** ********* *******, *** Shenzhen ****** ********* ** CGTN, *********, *** **** Russian ***** ******************** ***** ** ********-******************, ********** ** *** ********** '********** ********** ******' solution *** ******** ******* Police:

*********************** "**** **** *** media" *** ******* **** particular ****** ** ******** since *** ********* ** 2017. ** **** ****** as ******** ******* *** police *********** *** *************, such ** *********, *****, and ******:

***************** ** *** *********** **’* ****** “**** of ********” ** ******* during ******* ****** ** financing, **** ******** **** it **** ** * unicorn (****** ** $* billion ** ****) ** 2018. *******, ** ***** not **** *** ******* online ***** ************* ******** a *******, *** ***** are ** ******* ********** its ******** ********* ** revenue *******. *************’********* ************ *** ** *** PRC, ********* *** ***** banks.

*************

*********************** ***** **** ****** of *** ** ********, "the ********** ***** *** dropped **** *,***+ ***** / *** **** ** 80+ / ***".

Vague *****/******** ******

************* ****** **** ** can ****** * ****** face ****** * *** of **** ** ******* faces ** ***** *.* seconds:

************* ****** ** ** "accurate" ** * **** variety ** *********, *******, no ***** ******** *********** are *****:

*** *****/******** ****** *** vague. **** ***** **** no ******* ** ************* on *** **** *********** tests *** ******, *** ** ********** agency **** ******* *********** *** ********.

*******

*** ******* ** ****** captured ** *************'* ****** can ** ***** ** an******** ********* * ******** ****** webpage (***** *** ***** been ***** ****).

*** *** ****** *** high *******, **** ******* angles *** ********** **** pixel *******, **** ** the ******* ***** ************:

Implementation ********

*** ******* **** ***** such **** *********** ** scale, ********** ** **** time, *** ***********. ****** China’s********-**-****, ***** ** *:* verification (*.*. ******** *** person’s ***** ******* *******), this ** *:* *********** - ******** *** ******’* face ******* ********* ** millions ** ******. **** makes ******** * *********** issue, ********** *********** ******* such ** ****** *****, weather/lighting **********, ***.

**** *** **** ******** of ******** **** *** systems:

  • *** **** ****** ***** is * ***** ** a **** ** * bus ************* ***************** * **** ****** jaywalking, **** **** ** the **** ** ******. The **** **** ***** on ******* ****** *****: one******* ************ “**** ** ******** by *** *** **** negative ***** **** *** happened ** ****** *********** and *** ********** **”.

  • ********** **** ****** ******* ******* *********** ********* expert ****** *** ********** had “***********” ******** *** that “** ****” *** do ***% ****-**** ********* identification. * ***** ******** was ****** ******* *********** should “***** ******, **** release” *** *****.
  • ******* ***************** ******** **** ********* with *** ******* ************ of *** '**** *** shame' ****** ** *** system. *** ******* ** featured ******* ***** *** Shenzhen ****** **** ****** jaywalker's ****** (***.***.**/*****/) ** no ****** ******.
  • ******* ***** ************ ****** ** "**** than **%" ********. *** given **** *** ************ likely ********* ******** ** thousands ** ******, ***** positives *** * *********** risk.

*****/*********

***** ** *** ** the **** *********, *** particular ******** ************* ****** installed ** **** ** far **** *** **** one ***** ********* *********** in *****:

  • ****** *** ***** ** 2017, *** **** *********** ****** *********** ** catch *,*** *********** *** non-motorized ******* ******* (**** as ********), ****** ***** ************.
  • *** **** ******************* * '***** ** shame' ** **** **** to ***** ********** ************** **** * ********* logo ** *** ****** right.***** ***** ******** **** "******* *** facial ******** ** *** captured ********* **** *** demographic *********** ******** ** real ****".

  • ******** ***** * ***** face *** ****** ** "name *** *****" ********** in **** ***** ****** 300 ******,*** ********. ***** *********** **************** ******** *********** *** expanding *** ****** ** new **********, *** ******* these ******** *** ******* to *** ***** ** unclear.
  • ********** ***** ********* *** ** *** city ** **'** **** fall.
  • ********** ***** ********* * '***** ** shame' ** *** **** of ***** ** ****:

***** ***** ** ******** that ********* *********** ** expanding ****** ******* ******, such ******* *** *** omnipresent *** ** ***** no ******** ** *** national-level *******. ** ** absent ** ****** ****** in **** ***** ******, while ****** **** **** removed: **** ****, ************* **** ************ "* **** ** identify **********" ** *****'* third ******* **** ** Chongqing, *** *********.

**********

*** ********* ******* ** emphasizing ** ****** **** jaywalking ** * *******. The '**** ** *****' boards ********* **** ****** should *** *******. ** the ****** **** **** conditions ****** ** ******* less, ** *******, ******* or *** ***** ** a ****** *********** ****** present, *** ** ********.

*******, ***** ****** *********** to ***** ********** **** have *********** ******** / privacy ******. ***** **** this ****** ** *** nationwide *** **** *** issues *** ****** ********* in *****'********-******** *****, **** ** ****** overestimated ** ** ******* of *** ***'* **** '********', ****** ** ******** continued ******** ** *** how ***** ******* *******.

Comments (8)

**** ****** * ****, this ****** ** **** in *** ***** ******. I ***** *** ** us **** **** ****** a *** ******** ***** because *********** ******** ** cross **** **’* **** red!

Agree: 3
Disagree: 6
Informative
Unhelpful
Funny

********** *** ******* ***** are *****...

Agree: 3
Disagree
Informative
Unhelpful
Funny

**********: *** *** *** industry ******** ******** *** road

*** *** ** *********** article ** **** *****.  The ******* ** ********** is ***** ** *** assumption **** ***** *** primarily *** ****.  ***** is ********* * ***** safety *****; *** *** assumption **** *** ***** modes ** ********** *** secondary ********* ** *** road ** * ********* of *** **** ********.

** *** ***** ** facial *********** ** **** jay-walkers; ** ***** **** just *** **** ******** to *** ******* ******* that *** ********** ** watching.  ********** *** ***** to ******. 

 

Agree: 5
Disagree
Informative: 1
Unhelpful
Funny

** * *** **** as ******** *** ** the ** ** **** across **** ***** (***** is *********** ******* *** researched) ** * ******** reminder ** *** **** of ******* *** *** threats ** **** **** in *** ******.

**** **** ****, ** matter *** **** ******** lends ****** ** ***** and ******. ***** ******** and *********** *** **** maintained ** * ******** separation; * **** ** increasingly ********* ** ****** objective.

Agree: 9
Disagree: 1
Informative
Unhelpful: 1
Funny

**** ** **** ***** red ***** ******* **?

Agree
Disagree
Informative
Unhelpful: 1
Funny

*’* *** ****** **** the **** ****** *** this ****** ** ** reduce **********.

*** **** **** ********* have “******** * **** for ******* ************** **** as ***********...” *** **** crossing “********* ***** * bit **** ******** *** in ***** ** * firing *****,” ****** ** deterrent ******.

*** ********** *** *** ashamed ******.  ******** **** them.

**, **’* ****** * plausible ******* ***

*)******* *** **** ***

*)******** ***** ******** 

*)******* *** ****** **** their ***********

Agree: 4
Disagree: 1
Informative
Unhelpful
Funny

******: ***** *** **** controversy ** *** *** city ************** ********'* ***** **** displayed (** **** ******) on * "***** ** shame" ** * ***** intersection (*** ********* *****).***** ***** *********** ******* **** *** appropriate, *** *** ****** **********'* ******** was **** ****/****** *** the **** *********.

 

Agree
Disagree
Informative
Unhelpful
Funny

***** ***** ***** ****** whether **** *** ***********, but *** ****** **********'* response *** **** ****/****** get *** **** *********.

*** ****** ***** ** doubt, *** ******* **** sacrifices **** ** **** to *** ******* ** this ******* ******** ******** known ** “**********” ;)

Agree
Disagree
Informative
Unhelpful
Funny
Read this IPVM report for free.

This article is part of IPVM's 6,894 reports, 921 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports