81% of the 'suspects' found by the live face recognition system are innocent
by
A live face recognition system (LFR), used by the London Metropolitan Police Department, to identify suspects by comparing faces in databases of surveillance cameras and faces in a database, says ' very inaccurate accuracy and operation is opaque '. I have received an issue. At the request of the London Metropolitan Police Department, a survey on LFR has been conducted, highlighting the inaccuracy of 81% of those identified as 'suspect'.
met police live facial recognition trial concerns | University of Essex
https://www.essex.ac.uk/news/2019/07/03/met-police-live-facial-recognition-trial-concerns
81% of 'suspects' flagged by Met's police facial recognition technology innocent, independent report says | Science & Tech News | Sky News
LFR is also called Automatic Face Recognition System (AFR) and has been used in other events since it was first introduced at Notting Hill Carnival in August 2016. However, in the finals of the UEFA Champions League in 2017, more than 2000 football fans have been falsely judged as criminal suspects, and its accuracy has been questioned for some time.
In response, the London Metropolitan Police Agency, which operates the system, requested Essex University to investigate. Prof. Pete Fussie and Dr. Dollar Murray conducted an investigation and submitted an independent report.
(PDF file) Report on the London Metropolitan Police Service's Trial of Live Facial Recognition Technology
The report points out that there is a high possibility that the police will be unlawful to deploy LFRs if the court complains, as there is no explicit permission to use LFRs in domestic law.
Also, as a specific problem, in the cases under investigation, the number of suspects extracted by LFR was 42, but 8 cases were really criminals and 34 were innocent I have an issue. This indicates that LFR's recognition of suspects was only 19%. On the other hand, the error rate was “0.1%” according to the metric used by the London Metropolitan Police Department.
In the first place, the criteria to be included in the list of persons to be monitored were not clear, and the people who were trying to identify with LFR were also in different categories. The list itself is not accurate either, and there have been cases where a person who has already been tried is listed.
Operation itself has also been pointed out, Silky Carlos of the Big Brother Watch anti-surveillance campaign group told Sky News 'The question is when the London Metropolitan Police Department decides to discontinue LFR use.' Say.
'On the other hand, Duncan Ball of the London Metropolitan Police Department, who operates LFR,' is disappointed that the report is a negative and disproportionate tone. We believe that the public is looking for innovative ways. '
Related Posts:
