UK Facial Recognition Essex Errors Report

By: John Honovich, Published on Jul 05, 2019

Facial recognition trials in the UK have generated significant controversy and debate over the past few years. This week, it flared again when Sky News / Guardian trumpeted that 81% of 'suspects' flagged were false. However, their posts did not analyze the details nor share the original research report. IPVM has.

london MPS police face rec report

In this report, we share and examine the 128-page report including:

  • The new and enhanced methodology that Essex used to analyze these results from more recent facial recognition trials in 2018 and 2019.
  • Why police / human review of matches was so critical and so difficult
  • How the 'accuracy' statistics vary and what the tradeoffs in different approaches were
  • How logs recorded show logistical challenges
  • How UK / EU / GDPR rules for consent complicate the challenges involved
  • China / PRC facial recognition systems contrasted

****** *********** ****** ** the ** **** ********* significant *********** *** ****** over *** **** *** years. **** ****, ** flared ***** ******* **** / ******** trumpeted **** **%** '********' ******* **** false. *******, ***** ***** did *** ******* *** details *** ***** *** original ******** ******. **** has.

london MPS police face rec report

** **** ******, ** share *** ******* *** 128-page ****** *********:

  • *** *** *** ******** methodology **** ***** **** to ******* ***** ******* from **** ****** ****** recognition ****** ** **** and ****.
  • *** ****** / ***** ****** of ******* *** ** critical *** ** *********
  • *** *** '********' ********** vary *** **** *** tradeoffs ** ********* ********** were
  • *** **** ******** **** logistical **********
  • *** ** / ** / **** ***** *** consent ********** *** ********** involved
  • ***** / *** ****** recognition ******* **********

[***************]

Report ********

************* ** ***** ***** Rights ****** *** ******** a ***-**** ******:

Observing ******

*** **** ********* ********** of **** ***** *** here ** ** ********** observe *** ****** ****** recognition *********** ** **** they ***** *** **** happened **** *** ****** were ******* *** ****** recognition *** ********** *******. The ***** **** ***** report ***** *** * trials **** ********:

Problems ********* **** *********** *******

*** ******* ******* *** perhaps **** ********** ******* of *** ****** ** that ********* *** ***** difficulty ******** ********* *******, i.e., ** *** ****** on *** ****** *** same ****** ** *** person *** ****** *********** system **** ********?

** ***** **** *** 6 ******, *** ****** were *** **** ******* the ***** *** *** person ** *** ****** so **** ******* *** person, *** **** ** 8 ** ***** ** times *** *** ****** flagged ** *** ****** the ****** ****** ** the ******, ** *** chart ***** *****:

***** ********* **** *** unreliability ** ****** ********* the ****** ** * key *******:

Sky **** **% *********

*** **% ******** ** Sky **** ******** ** matches **** *** ****** were **** ** ******* dismiss ** ***** ***** matches (*.*., *** ****-**** surveillance ***** *** ******* much ********* **** *** watchlist *****):

***********, *** **% ** wrong ** ***** **** only ** ******* **** were **** ** ** verified *** ** *****, 30 **** ******** ** being *****, *** * rate ** **.*%. *** difference ** **** * matches **** ****** ** be ********. ****** ***, the **** ** '****'.

Debate **** **********

**** *********** * ****** on *** ** ******* the *********** ** ***** systems.

************* ****** ******** **** use *** ***** ******* of ***** ******* ****** them ********** ** **% or ******. *** ****** mentions **** ********:

** **** ****, *** system ****** ******* ********* of ***** ** ****** walking ** **** ***** 6 ****** *** **** ** **** ****** ** be *****. ********** ** this *** (*** ** false / **,*** ***** scanned) ***** * ***** positive **** ** **** 0.3% ** ** '********' of **.*%).

***** ** *****? **** are, ** **** *****. The ******** ******* ***** is **** ********.

Hard ** Visually ******

*** ******* ** **** it ** **** ** visually ****** ******* * match ** ***** ** wrong. ****** **** *****. For *******, ** ********* *********** ****, **** had ** ******** '*************'******* *** ****** *** generating *** **** ******, e.g.:

*** **** *** ********* high ******* ***** ******, not ************ ***** ***** where ******, ********, ******** variances, ***. *** **** it **** **** ** know *** **** ** just ******* ** * screen.

*************, *** ***** *** not ******* *** ******** of *** ******* ** show *** **** ** was ** **** ******* the ********* ***** *** the ****-**** *****.

Logs **** **********

*** *********** ******** * detailed *** ** ************ for *** ** *** test ***********. ** ******** one ******* **** ** front ** * ******* being ********:

** **** ******** ******** mistakes. ******* *** **** interesting ** * ** them ********* **** * minutes. ** *** *****, the ********* ******** ***** whether * ***** ** accurate *** *** ****** one ***** ***** *** ready ** ****** ***** they **** *** ******** processing *** *****:

*** *********** ******* ** that **** ** ***** have ********** *** ********** scores (** *.** *** 0.60). *** ****** ***** that *** **** *********** vendor (***) *** * threshold ** *.**:

*** ******* ********** ***** can ** ********, *** higher ** **, *** more ****** *** ****** rejects ***** ******* *** misses **** ******* *** vice *****. ** **** not ****** ***'* ****** recognition ** ** ****** opine ** *********** ******** but ******* **** ***** would ********* ****** ***** matches *** *** ******* "81%" ***** ****.

Technology ************

*** ****** ***** **** the **** ************ *** the ******* **** ********:

** ******* **** ******** about **** ******* **** used.

Positioning *** *******

*** ******* **** ********* mounted ********, ** ****, as *** ***** ***** from *** ****** *****:

*** *********** ******** ******* is **** *** *** strove ** ******* ******* and ***** ********* **** facial *********** *** ***** performed, ** **** ******* notes:

** *** **** ****, some ****** **** ******** after ******** *** ****** recognition:

China / *** **********

*** ******'* ******** ** China (***) ****** ** such ****** ******* ** facial *********** *** **** it **** *** ******* of *********** ******** ** it ** ************* ********** to ******** *** *********** of ***** *******.

**** ** **** ************** is *** **** ***** PRC ****** *********** ******* work (****** ***** **** been ******* ********* ** the ********, *.*.************************* ***********).

*** ******* **** *** PRC ****** *********** ******* is **** **** **** more ***** ***** ***** on ****, **** ******** data, **** ********* ****** (i.e., ******** ******** **), and **** (** **) legal ************. 

*** ******* **** ** that ****** **** *******, especially **** *** *** dealing **** ***** ********** (e.g., ****** **** *** a *** ******* ** a *** ********, ***** PRC ******* ****** ***** have ********). ** ** inevitable **** ***** *** going ** ** ******** mistakes **** **** **** humans ** ******** ****** if **** **** ***** accuracy ** ********.

**********, ***** **** **** and ******* *** ****** about ***** *** ******* performance, ** **** ** hard ** ****** **** how **** ** *****.

UK ****** *********** ******?

***** ******* **** ********** about *** ****** ** facial *********** ** *** UK. ***** *** **** to *** **** (*******, software, *********, ***.) *** the **** ** ******* informed *******, ** ** hard ** *** **** as ***** * '**********' or *********** *******. ***** are ****** *** *** people ***** ******** *** the **** *** ********** involved.

**** ********* *** ****** as *** ********** ******** but *** ********** ** visually ********* ****** *** likely ** ************* ****** given *** **** ****** look ***** ********** ***** the *********** ** ***** quality ** ********** *******, open ****** ***********.

Comments (3)

*********** ******, ***** ***. It's ****** ** *** real **** ***** *******.

******** ******* ****** ***** NEC **** **** ********* by ********** ********** ** Justice *****://***.****.***/*****/*******/******/**********-*****-******-********-*******-**-*************-******-***********-********

** *** ***** **** NN ******* ** ****** face ******** (*****, *** and ****..) ******** ****** in **** **** ********, where ******* ***/*** ******* are **** *** **** position ** *** ** perfect ** ** ********.

** ********** *** "******" (we *** ** ****** to **** *********) ** has **** ******** **** recognition ********** *** ************ performed **** ** ******* NIST *****. 

************* ****** ******** **** use *** ***** ******* of ***** ******* ****** them ********** ** **% or ******....

** **** ****, *** system ****** ******* ********* of ***** ** ****** walking ** **** ***** 6 ****** *** **** 30 **** ****** ** be *****. ********** ** this *** (*** ** false / **,*** ***** scanned) ***** * ***** positive **** ** **** 0.3% ** ** '********' of **.*%).

** *** ************* ** any *** **** ***** negatives **** ******* ** their **********? *** *** this ** ********* **** sounds ****** *****.

** *********, ************* ********* just **** *** ***** numbers, *.*.,**** ** * ********* explanation:

** '*** *** **%' is ****** *****. **** does ** *** **%? How **** ***** *** it **, ***.?

***** *** ********* **** to **** ***** ********* into *******, ** ***** is ** ******** ******** between ***** ********* *** false *********.

* ********** ********* *** ******** ***** to *********** ****, *.*.:

Image result for roc curve facial recognition

*******, ** ************ *********, they *** ******, ** ever *****.

 

Login to read this IPVM report.
Why do I need to log in?
IPVM conducts unique testing and research funded by member's payments enabling us to offer the most independent, accurate and in-depth information.

Most Recent Industry Reports

Hikvision Global News Reports Directory on Oct 17, 2019
Hikvision has received the most global news reporting of any video surveillance company, ever, ranging from the WSJ, the Financial Times, Reuters,...
Camera Calculator V3.1 Release Improves User Experience on Oct 17, 2019
IPVM has released a new version of our Camera Calculator, V3.1, with significant user experience improvements, a new development plan, and an...
Securing Access Control Installations Tutorial on Oct 17, 2019
The physical security of access control components is critical to ensuring that a facility is truly secure. Otherwise, the entire system can be...
Access Control Course Fall 2019 - Last Chance on Oct 17, 2019
Register Now - Fall 2019 Access Control Course. Thursday, October 17th is the last day to register. IPVM offers the most comprehensive access...
US DoD Comments on Huawei, Hikvision, Dahua Cyber Security Concerns on Oct 16, 2019
A senior DoD official said the US is "concerned" with the cybersecurity of Hikvision, Dahua, and Huawei due to "CCP" (China Communist Party)...
Pelco Sarix Pro3 Camera Tested on Oct 16, 2019
Pelco has released their Sarix Professional Series 3 cameras, claiming "more security detail in challenging scenes with excellent low light and...
IPVM Camera Calculator User Manual / Guide on Oct 16, 2019
Learn how to use the IPVM Camera Calculator. The guide below includes instructions, images, gifs, and videos demonstrating and explaining the...
Altronix Claims Tango 'Eliminates Electricians' on Oct 15, 2019
Power supply provider Altronix claims its new Tango power supply 'eliminates the need for an electrician, dedicated conduit and wire runs'. In...
Pelco CEO Out, New CEO Found on Oct 15, 2019
Just 2 months after Pelco was sold, Pelco's CEO is out, with Pelco bringing in an outside President and searching for a new CEO from the industry,...