30 Million Criminal Face Database Tested (Captis Intelligence)

By Sean Patton, Published Apr 27, 2020, 09:36am EDT (Research)

** ******* ******** ******** *** *** available *** ****** *********** ******** ** subscribing ******** ************.

IPVM Image

****, **** ******* ************ **** ********* ********.

** ********** ** ******' *******, ********* to ********* ********, ** *** *** well ** ******, *********:

  • *** ******** ** *** **** ***********?
  • **** ***** *** ******** ** *** database?
  • **** **** ******* **** ****** ********?
  • **** ******** **** **** ****** *******?
  • **** ****** **** *****?
  • *** **** ** ******* ** ***** facial *********** *******?

Executive *******

IPVM Image***** *** *********** ****** **** *** the ****** *********** ******** ******* **** look ******* ** ******** ******** ** their ********, *** **** ******* ******* was **** ** *** **** *** us ** ****** ******* ***** ******* were ******* ** ****** *************. *** risk, *********, ** **** ****** ***** be ************* ** ********* *********.

***********, ******, *** ****** ***** **** in ********* ****** ******* *******, ********** quickly ****** ******* *** ****** **** from **** *********'* ***.

*******, ********* *********** ******* **** **-***** criminal ********, ** **** ****** *** watch *****, *** * ***-****** $** per ***** **** ******* ********* **** would ** ***** ****** *** **** users (****** *** ******* **** ****** pricing ** **********).

Test ********

** *** *****, ***** **** ******* positives:

  • ***** **** ********** ********:**** ******** ********* ********, **** **%+ confidence ******* ** ******* ******* ***** on ***** ******.
  • ***** ****** *******:****** ******** ****** ******* ****** * seconds *******, ********* ******, ****** ****, and ********* **************.
  • **** ******** *******:******* ******** **** **** ********** ******* for **** ******* ******** *****. *******, typical *** **** **** *********** *******, detection *** ***** ********** ******** *** to ****, **********, *****, ***** *** commonly **** ** ********* ********** ** avoid **************.
  • ******** ********** ***********:******* ******* ******** ********** *********** ** the *******, ********* ******, ***/**** ** birth, *** ******** ******** ***********. *******, the ******* *** ************ ***** ** what ***** ************* ******** "******" ***********.
  • **** **** ***** *********:**** ******** ********* **** ***** ***** IP ************ ******* *** ******** ********* body-worn ****** **********.

*******, ** **** ***** *********:

  • ** ***** ******* ** ***** *********:**** ********** ** ***** ***** ********* criminals, ****** ****** ** ****** ***** matches. ** **** *** **** ** verify ******* ** *** *** ****** of ********* ** **** **** (** were ***) ** ***** ********.
  • ************* ** ******* *** ******** ******:**** ********** ** ***** ***** ** a ****** ******, ***** ******* ***** change ***** ** ********* ******* ** by ********* ** ********* **** *-* frames ** *****.
  • ******** ***** ********:** ***** ********* ****** ******, *** obscured **** ******** ******. ****** **** that **** ** *** ***** ********* records ** **** ** ***** ** million ***** ****, *** ******** **** they ******* ********* ******* ** *** records **** ********* ********** *******, ******* that ************* **** ********* *********** *** what *********** ** ********.
  • *** *********** *******:****** ** ********* **** ********** **** Milestone ********. ****** ***** **** *** open ** *********** **** ***** ***** but ** *** **** *** ***** integrations ** ***********.
  • ********* ************ *****:****** ***** $** *** ****** *** month, ***** ** ************* **** ********* than **** ****** *********** ****, *** does *** ******* ****-**** ********. ****** said **** ** *********** *******, *** they *** *********** ***** ******* ******.

Post-Incident ****** ********

** ****** ******' **** (** ****** Subject *************) ******* ********** ******* ************ ******. ** ** * ********** ****** plugin, *** **** ** *********'* **** Centralized ****** ********. ****** ********** **** ODSI ** * ****-********, ************* *******:

** * ****-******** ************* ****, ***** is *** * **** ****** *********** scanning ********* ** *** ********** ********** of ********* **** ** *** ******* public. **** ** * ***-********* ****-******** Investigation **** **** ** * ** million+ ******** ** ******** ********* ******** images/files.

****** **** **** ** *** **** to ***** * ****/******** ******* ** their *******.

Privacy **** / ******* ***********

****** **** **** **** ** *** return ******* ** ********* ********* **** States ***** ******* **** ******* ***** use (*.*. ********). ********, **** **** they **** * ******** ********, *** wanting ** ** ******* ** ***** or ********.

**** **** **** ******* ****** *** added *** ******* ***** ** *****, as ******** *******, ********* ********, *** updated. *******, **** ****** **** **** jurisdictions *** **** ** *** ** remove ******* **** ***** ********** *********.

Captis ****** ******** *****

*** ********* **-****** ***** ******** ******, including **** *** **** ** *** facial ***********, ******** *******, ***** *******, and ***********:

High ********** ******* ** ******* *****

** *** *****, ***** ******* ** high ********** (**+%) ************ **** ******** similar ** *** ******** ****:

IPVM Image

*******, ********* ********* **** *** ******* was ******* *****, ****, **********, ** was ******** ******** ************** ** *******:

IPVM Image

**** ***** ************* ******** *** ***** of ****** **** ************* ******, ** many ********* ** ***** ****** *** want ** ***** ********* **** ***** for ***** **** ******* **** ***** it ****** *** ****** ** ****** recognition ******* ** ******** ****.

Face ******* / ************** ********

****** ******* **** ******* ********* ***, gender, ****** ****, *** *********, ***** we ***** ******** ******** *******. ************, image ******* *********** *** **** **** is **** ********, ***** ** ** not ******** ******** ** ****** *********** systems:

IPVM Image

** ***** **** ******** ** ** generally ******** ****** ***** *** *** case ** ********* ******** *** ** one, ***** ********* *** ****** ********* gender *** ********* ******* *** ****** telling ****.

Quick ***** *******

*** ******* ** ****** *** **** matches ** ****** **** ******** * clicks *** ******* *** ******** ****** 5 ******* ** ******** *** "**** Image ** ******" ****** ** *** Milestone ******:

***** *** ***** ******* ***** *** 30 ******* ******** **** ****** ** searching.

Duplication ** **** ******

*** ******* ****** ********* ******* ** many ****** *******. *** ********* ******* had *** **** *****, *******, ****** noted **** **** ************* ******** ********* details ***** * ********* ******** *** that ***** ********* ******* **** **** varying ********** *** ******** ******* *******:

IPVM Image

****** **** ***** **** **** ***** their ******** *** ********* ***** ******* and ****** ***** **** **** *** same ******** ********** *******.

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

** ***** ** ******** ******** ** known ********* ********* **** ****** *** not ****** ******** **** ******* ** that ********** ****** *** ***** **-** results:

**** ** ** **** *** ** the **** ** *** ********* **** database, ***** ***** **** *** ***** person ******** ***** *** ******** ** potentially ********* ** ******* **** *** visibly *******.

****** ******** **** **** ****-*******/********* ********* criminal ******** *** *** ********* ** their ******* ********* **** **** *** enforcement ** ******* ***** ******** ******* jurisdictions **** ** *** ***** **** with ********** ********** (**** ******).

Body **** ****** ***** *********

** ***** **** ****** ******** **** confidence **** ******* ** ***** **** body-worn *******:

IPVM Image

** **** *****, *** *** ***** of ********* *** **** ******* ********* the ********** ** **** ********* *** provided ****** ********** ***** **** **** ceiling-mounted ************ ****** ********.

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

****** **** *** ***** ******* ** narrow **** **** ***** *******, *** only ******** ******** ****** ******* **** highest ** ****** ********** ** **** versa, ***** ** *** * ******* feature.

****** **** **** ********* ** *** add ******* ** "*********" ****** ********, instead, **** **** ** ******* ** much *********** ** ******** *** *************, to ***** * ***** ** **** the ********:

*** ****-******** ******* ****** ******* ********** investigation. **** ** * ****-******** ****, allowing ******* ************* ** *** ******* to *******/**** ********.

*******

****** **** ********* **** *** ********* plugin ***** $** *** ****** *** month. **** ** ************* **** ********* than **** ****** *********** **** *********, which ********* **** $**-$** *** ****** per *****. ********, ***** ****** *** only ** **** *** ****-***** *************, most ********-******* ****** *********** ******* ***** real-time *******/******** (********* ********** * ******* price *** **** ********** *** **** increased ******** / ******** *****).

****** **** **** ** ***** *********** pricing *** **** ***** ***** *** quantity ********* *********:

********* *** ********* *** ******** ************ starting ** ** ** **** *******, customers *** **** ** ***** ****** or ***** ***** ********** ****** *** pricing. * *****-***** ******* ****** ** also *********

VMS *********** - *********

****** ** **** ********** **** ********* XProtect. ************, ****** **** **** ** not **** ***** ************ ** ***** roadmap, *** **** ***** ******** ***** integrations ** *********:

** *** **** ** *********** **** other ********** ********

*************, ***** **** ** ****** ****, users *** ****** * ******** ***** from ***** *** *** **** ****** to ******' *** *******.

*******

** ****** ** *** ****** *** use ** ********* **** ****** ** expand, ** * ***** ********** ** typical ****** *********** ******* ** **** each **** *** ** ***** ***** own ********, ***** ** ****-********* *** inherently *******. *** ******* ** ****** tens ** ******** ** ****** ************ will, *********, ** ********* ** *****.

*******, ******** ***** *** ** ******* such ******** *** ****** **** ******* entities ** *** ******* ****** ** innocent ****** ** ***** * ********* criminal **** ** ** ********* *** could ****** * ***** ******* **** a *****, ******* *** ****** ********* perspective.

Voting / ****

Comments (4)

I can see there being a lot of potential problems for the database here. What happens to the facial image of someone who has "served their time", there remain as an active criminal? Also, depending upon the other data contained, how does this pass some of the privacy laws of California or Illinois for example?

Agree
Disagree
Informative: 4
Unhelpful
Funny

Thanks for the questions Malcolm, I added a section to address these concerns:

Privacy Laws / Mugshot Maintenance

Captis said that they do not return matches of convicted criminals from States where privacy laws prevent their use (e.g. Illinois). Moreover, they said they take a cautious approach, not wanting to be subject to fines or lawsuits.

They said that mugshot images and added and removed every 24 hours, as criminal records, including mugshots, are updated. However, they report that many jurisdictions are slow to add or remove records from their electronic databases.

Agree
Disagree
Informative: 2
Unhelpful
Funny

in addition, what if they have changed their appearance? The mug shot shows someone with a bald head but what if they have since grown out their hair? Have a full beard?

Agree
Disagree
Informative
Unhelpful
Funny

Change in appearance does not have a major effect on facial recognition. Almost all FR engines can match a scanned picture to the enrollment picture with many years of age separation between the pictures, including change of appearance due to age, hair, etc. At the end of the day, all post-incident facial recognition results involve a human making the decision on whether or not the match results are a positive identification.

Agree
Disagree
Informative: 2
Unhelpful
Funny
Subscribe to IPVM Research to read the full report.
Why do I need to subscribe?
The IPVM Research Service includes products tests and shootouts plus competitive and financial analysis, helping decision-makers better evaluate purchasing, partnering, developing, and/or competing against companies in physical security.
Already have an account?
Loading Related Reports