Japanese Stores Facial Recognition Sharing

By: Carlton Purvis, Published on Apr 21, 2014

A controversy is brewing in Japan after a newspaper was tipped off about a practice stores use to combat shoplifting. A network of stores uses facial recognition technology to tag customers as shoplifters and relay that information other stores. In this post we examine the practice and compare it to similar U.S. efforts.

* *********** ** ******* in ***** ***** * newspaper *** ****** *** about * ******** ****** use ** ****** ***********. A ******* ** ****** uses ****** *********** ********** to *** ********* ** shoplifters *** ***** **** information ***** ******. ** this **** ** ******* the ******** *** ******* it ** ******* *.*. efforts.

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

**********

***** ******** ************ ******* ************ ****** ** *** Tokyo ************ **** **** installed ****** *********** ******* to ******* ****** ** shoplifters ** ********* ***** for **********. ***** ****** are ******** **** * database *** *** ****** in *** ******* *** also **** ****** *********** systems. *** ****** ***** store ******** ****** **** a ********** ** ******* customer ***** **** *** store.

*** ******* ******* ***** face **** ***** ** a *****, *** **** only *** ***** ****** by ******** ** *** database. **** *** *** database ****** ** “*******” by ****** ** *** network.

Japan ******* ********

*** ***** **** *** Japanese ********* **** **** under *** ******** *********** Protection ********* ******** ** ******** are ********** ******** *********** which ** *** ** be ********** ***** *******. ******* the ****** ***** ****** may ** * ********* of **** ***.

* ****** *** ********* spoke ** **** *** practice ** ********** ******* it ***** ***** ******** to *** ******* ******** or ** ********** **** visiting ****** **** **** never **** ** *** if ******* ** *********** added ** *** ******** there ** ** *** for **** ** ******.

***** *** **** ********* issues **** ********. **** what *** ***** ****, these ******** *** ***** taken **** ************ ******* and ******** ** ***** images ***** **** ************ cameras -- *** ************ camera ****** ***** ******** to ****-** ****** **** driver's ******** ** ******** photos *** *******.

Sharing ********** ****** ** *** *.*.

** ********, *** ******** of ******* ****** ***** retail ****** ** ****** in *** ****** ******. It’s *** ** *** form ** ****** *********** technology, *** ***** ** an *********** ******* ****** for *** ***** ********* where **** *** ***** video, ****** *** ************ of *********** **** *** created ** **** ****** organized ****** *****.

* **** *************** ******** form ** **** *** been ****** ** *** U.S. *** * *****: Stores ************ ********* ** images ** *********** ** other ********* ** *** area.

******* ********** ** **** in *** *.*., **** someone **** ****** *********** bigger ****** **** **** a ***** ** **** when ****'** ****** *** add **** ** * company-wide **** ******** ** paste **** ** ******** boards. ** *** **** of *** ******** *******, people ***'* **** ***** photos *** ***** *********.

Hard ** **** **** ****** *** ***** *** **********

** *** **** ********* ** find *** ***** ****** use *** **********. *** story **** *** *** what ******* ***** *** technology ** **** *** of *** ****** ***** it ** **** ********* several ************ (**** *** and *****) ** *****. None ** **** **** they **** ***** **** software ** ******** ** answer. *** *** ***** says *** ****** ***** 50 ********* ********* *** using *** **********.

******** **** ********** ***** as **** ** *** manufacturer ****** ** *** story, ** *** ***** it ** ********* ******* to ***** ***** ****** and ** *********** *** people ** ***********?

Comments (5)

Suppose that magically (1) only shoplifters and no others were identified as risks and included in the database, and (2) erroneous matches never occurred. If this were the case, then it seems not unreasonable that high risk individuals should be closely monitored to mitigate losses.

On the other hand, suppose that (1) innocents are often erroneously or even maliciously identified as shoplifters and included in the database, and/or (2) erroneous database matches often mis-identify unrelated shoppers as risks. If this were the case, then the approach seems unreasonable and also counter-productive in terms of allocation of scarce retailer resources.

If the truth lies somewhere in between, at a minimum, the inability of databased suspects to understand or address systemic errors seems problematic. What indignities is an innocent likely to be subjected to, beyond closer scrutiny while shopping? Will these databases make it difficult for wrongly accused to get a job or a clearance because security investigations and pre-employment screenings use information from these databases? At a minimum, isn't it likely that job seekers at those retail facilities will walk through a monitored area and that management will be notified if they are flagged as a high shoplifting risk?

Also, I might be more sanguine if we didn't have the example of the do not fly list. I expect that in many cases it performed as intended, but the press indicated that some innocents were inconvenienced while growing pains were addressed, though I haven't heard of any issues in a while now. By its very nature, it seems the affected parties learned about and consequently had at least some form of recourse, although (for example) losing non-refundable tickets and non-refundable accommodation down-payments and then, three months later, having one's name removed from the list does not seem at all equitable. In the case of this retail shoplifters list, one could be even more powerfully negatively affected yet have no idea of the source of negative information. For example, an innocent person wrongly flagged as a suspect might never be called back after every employment interview, yet never understand the problem that needed to be addressed. Even if the source were discovered and even if the list were eventually corrected, would the reputation ever be rehabilitated? Would the losses suffered during the erroneous blacklisting ever be compensated? But more vigorous loss recovery options could increase resistance to ever acknowledging and correcting an error in the first place.

For a somewhat related issue, you might take a look at Spiceworks' recent discussions on issues associated with the way IP blacklisting is currently managed.

P.S. This was a fun article with interesting implications. Keep up the great work, IPVM!

I blieve that if a shopper is prepared to enter someones premises with the intention of helping him/herself to thier property, he forfeits the right to maintain the privacy of his personal identityl, period.

It doesn't say that the suspected thieves are subjected to any sort of mistreatment, right? It just alerts LP that a suspicious person is in your store and they should keep a close eye on them. If that's the case, no harm, no foul.

Now, if these suspected individuals are harrased or intercepted at the door, then you might have something to gripe about.

Most Privacy laws require the holder of personal information to advise the person that the holder has the information and how it is being used. But, if a facial image has been provided to the police as evidence of a crime, it would be inappropriate. and possibly unlawful, to advise the person in the photo that the holder or the police have the evidence.

Prviacy laws allow Personal information to be collected so long as there is a 'necessity' for the information. It is appropriate for a person or a company to protect themselves from crime by circulating the photo to staff, as well as warn other potential victims, otehr shops, of the threat. Best to use the term 'alleged' before the description of the criminal activity.

"Most Privacy laws require the holder of personal information to advise the person that the holder has the information and how it is being used."

This depends on the country. In the US, there's really no law regulating / stoping use of such information. In Europe and Commonwealth countries, the opposite tends to be true.

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.

Related Reports on Facial Recognition

Ireland National Children's Hospital Chooses Hikvision End-to-End With Facial Recognition on Dec 05, 2019
The world's most expensive hospital project ever, the New Children's Hospital in Ireland, has chosen an all-Hikvision surveillance system including...
Hikvision Global News Reports Directory on Dec 02, 2019
Hikvision has received the most global news reporting of any video surveillance company, ever, ranging from the WSJ, the Financial Times, Reuters,...
ISC East 2019 Show Report on Nov 21, 2019
IPVM has finished in New York City covering both days of the ISC East 2019 show. Here is a 6+ minute general walkthrough: Inside this report,...
Hikvision Dual Lens Face Recognition Camera Tested on Nov 19, 2019
Hikvision's Dual Lens Facial Recognition camera, claims that it "adopts advanced deep learning algorithm and powerful GPU to realize instant face...
Critiquing Carnegie's AI Surveillance Paper on Sep 25, 2019
The Carnegie Endowment has issued an ambitious paper on the Global Expansion of AI Surveillance. While its aim is applaudable, the paper has...
IFSEC 2019 Show Report on Jun 19, 2019
The UK's largest trade show, IFSEC, is underway and IPVM has been examining what is new and happening at the show. Inside, we cover: Huawei...
19 Facial Recognition Providers Profiled on Apr 23, 2019
IPVM interviewed 19 facial recognition providers at ISC West to understand their claimed accuracy, success and positioning. 9 from China, where...
Hikvision Chairman Warns Of AI New Entrants Victimizing Customers on Mar 12, 2019
Hikvision chairman Chen Zongnian has warned that customers will become “victims” of AI hype from new industry entrants, even saying such newcomers...
Massive Leak Of Chinese VMS Provider Exposes Xinjiang Surveillance on Feb 20, 2019
A subsidiary of China’s claimed largest VMS provider is tracking the precise location and ethnicity of millions in China’s Xinjiang region,...
Casino Surveillance Pro Interview: James Lathrop on Feb 15, 2019
James Lathrop [link no longer available] has been working in casinos for almost 25 years. During that time, he says he has held "just about every...

Most Recent Industry Reports

Video Surveillance 101 Course Opened on Dec 12, 2019
IPVM is adding a Video Surveillance 101 course, designed to help those new to the industry to quickly understand the most important terms,...
Verkada Notification Outage on Dec 12, 2019
Verkada is suffering an event notification outage and analytic search failures. Inside, we examine what the issues are, what Verkada told IPVM...
Hikvision DS 2nd Gen Intercom Tested on Dec 12, 2019
With its newest IP intercom, Hikvision proclaims users can 'get full control over an entrance' regardless of where it is installed, home or office...
Honeywell 30 Series Cameras Tested Vs Dahua and Hikvision on Dec 11, 2019
Honeywell has infamously OEMed Dahua and Hikvision for years, but now they have introduced an NDAA-compliant line, the 30 Series, claiming "lower...
"Good Market, Bad Business Models" - Residential Security on Dec 11, 2019
Industry banker John Mack, at his company's annual event, took aim squarely at the problems in the residential security...
IP Camera Browser Support: Who's Broken / Who Works on Dec 10, 2019
For many years, IP cameras depended on ActiveX control, whose security flaws have been known for more than a decade. The good news is that this is...
Acquisitions - Winners and Losers on Dec 10, 2019
Most major manufacturers have been acquired over the last decade. But which have been good deals or not? In this report, we analyze the...
IP Camera Installability Shootout 2019 - Avigilon, Axis, Bosch, Dahua, Hanwha, Hikvision, Uniview, Vivotek on Dec 09, 2019
What are the best and worst cameras to install? Which manufacturers make it the hardest or easiest to install their cameras? We tested 35 total...
Viisights Raises $10 Million, Behavior Analytics Company Profile on Dec 09, 2019
Viisights, an Israeli AI analytics startup marketing "Behavioral Understanding Systems", announced $10 million Series A funding. We spoke to...
Disruptor Wyze Releases Undisruptive Smartlock on Dec 06, 2019
While Wyze has disrupted the consumer IP camera market with ~$20 cameras, its entrance into smart locks is entirely undisruptive. We have...