Evolv Is Not A "Weapons Detector", It Is a Smarter Metal Detector

By Nikita Ermolaev, John Honovich, and Donald Maye, Published Jan 11, 2022, 09:44am EST (Info+)

While Evolv's CEO declared Evolv is "a weapons detector, not a metal detector", IPVM has determined that Evolv is a metal detector, though smarter than conventional ones, based on conversations with Evolv's executive team plus research into Evolv's patents and academic papers.

IPVM Image

Evolv cannot precisely detect what is truly a weapon, regularly mistaking what they admit are "benign" objects, including Chromebooks (as we recently analyzed) plus 3 other common false alert sources, as weapons.

In this note, we examine Evolv's technology, including how it operates, the systematic false positives it generates (from its own documentation), and how it is clearly smarter than conventional metal detectors though still limited in its intelligence about "weapons".

Not * ***** ********

***** ********, ******* *************, **** **** are "*** * ***** ********",** ***** ***, ** ********* *****:

IPVM Image

False ********* ***** - **** ******* *******

*******, *****'* **** ****** ****** ** false ****** ******* *********, *********, ******** cases, *** *******, ** *** ******* from*** *******'* **** ******** ********** *****:

IPVM Image

**** ***** * ****** ******* - objects **** *** ******** *** ****, as *** ************ ***** **********:

IPVM Image

***** ****** ********** *********** ******* ** actual ****** - **** * *** - *** ******* **** *** ******* not ******* (******** *****, *******, *********, umbrellas) ******* ***** ******* **** ***** characteristics **** **** ****** ***** *** Evolv **** *** **** ********** ************ to ************* ******* ****.

Conventional ** *****

**** ***** ******* *** ************ ***** Detector *** ****** ******** ********* *******, however, ***** ******* **** * **** sophisticated ******** ** ******** ******** ****.

***** ************ ***** ********* ****** ******* the ******** **** *** ******** ******, Evolv ******* ******** * *********** ******* of ***-********** *******, ********* *** ************** index ** ******** ******* ***** ******* and * ************** ******.

***** **** * ************** ****** ********* the ******** **** *** ******** ****** scanned ** * ***** **** *********. From ***** ***** **************** ****:

*** ************** ****** ... *** ** configured ** ********* * ****-*** ****** or ******* ***** *** ** ******* by * *** ** ********** ********* position, *****, ****-******, *** ************** *****. Best *** *** ** ********** ** a **** ******** **** ******** *** difference ******* *** ****** ************ *** those ********* ** * *****...

*** ****** ******* ******** **** **** the ****-*** ****** ** *****, * corresponding ***** ** *** ******** ************** Index *** *** ****** ** ********** and ****** **. **** ***** ************* the ******'* ******** ******** *** ** dependent ** *** *****, ****, *** electric *** ******** ********** ** *****. The ****** ***** ** ******** ** a ******* ** ************* ****** ********** to **** * ******/***-****** ********.

************ ***** ******* *** ******* ******** capabilities **** *** **** **** ******/***-****** discrimination, **** *** **** ******:

…* ******/***-****** ******** *** *** * classifier ******* ** ** * ******* learning *******, ** ***** **** ******* examples ** ******* *** ***-******* *** used ** ***** *** ******** ******* algorithm ** ********* ** * ***** detected ****** ****** ** ************* ** a ****** ** ***-******.

************ ***** ********* *** ****-************* **** for ****** **************** (*** ***** ****, or ***** *** ********* ** * received ******) *** ** *** *********** the ********.

*******, **** ****** ***** ** ******* than * ************ ***** ********, *** process ***** *** *********** ************* *********** in *************** ****** ******* **** ****** weapons. ***** **** * *********** ***** to ****** *** ****** * ******* object ** ** ***** * ******* of ***** *******. **** ** ***** on ******* ******** ********** ** ***** weapons (**** ** *** ************** *****). Nonetheless, ******* ****** ******* (**** ** umbrellas, *********, *******, ***.) ******* ******* properties ** *** ******* **** ***** uses.

** ******, *** ** **** ******* the ****** ** ******** ** *****'* process ** ******* ** ********* ******** objects *** ********* ****** ******* ***-******** weapons.

Value ** ******* ***** *********

****** ***** ********* ******* **** ******* potential ********, *.*., ******** *** ****** and ***** ** ******** ******* **** generate ***** ******.

*** *****, **** *** ************ ***********, detects ***** *** ***** *** *********** limitations ** ********* ******* ****** ******** objects *** ******* ***-******** *******. ************* this ** ********* ** ******** ******** both *** ******** **** ***** *** deliver *** **** *********** ** ***** are ******* **** *****'* ******.

Evolv ******** / *** *******

********** ** *** ******** ** ***** to **(***** *** *** ** ********), ** did ***** **** *****'* ********** ******* in ********, ***** *** ******* ************ that **** **** ****** ***** ******* and **** **** ***** ******** ***** alerts ** ***-*******. ** **** ***** Evolv ***** *** ********** ** **** Evolv ** ******** **** * ******* or ****** ***** ********. ** **** time, **** ***** ** **** **** time ** ***** ***** **** ***, thereafter, **** ******* ** ***** **** IPVM.

** ** ** **** *** ******** we ****** **, ** ** **** Evolv ***** ** ******* ******** ** input, ** **** *********** ****** *** reporting *** ***** ** **** *** public.

Comments (12)

******* ***** *** *****-******** ****** *** a ******, * ***** **** ** are ****** **** *** *** ****** of ** ** *******. ** *** last ** ***** ** **** **** from ** ***** **** ** * very ******** *** ******* **** **** processing ****** (** ******** ** *****) to ********* **** ** ********* ** be ********* ** * *********. ***, these *********** ** ** ***** ********** ready *** **** ** ***** **** complex ** *********** **** ********** ******** are ****** **********.

** ** *** *** ** *** point **** ** ** ********** ** problem **** **** ******* ************** **** what *** ********* ** **********. ** is *** *** ******** **** ***** "cheap" ** "****" ** "******" ** the ***** ** **** ***** **** to ***** ** ** *** ****** datastream (********* *****) *** ****** ** to **** ********** ******** *********.

* ***** ***** *** ****** *** a **** *********** *******. **** ****** time *** ****** * ********** ******* that **** **** *** ******** ** increase *** ******** ****** ************ ** metal ********* ** *** **** *** AI ******** ***** ********* ******* ***** 2010. * ***'* ***** **** *** going ** ** **** ** *** the ********** ** ******* ***** ********* reliability ** *** ***** ** ***** able ** ******* **** ****** *** cases, **** **** ******** ******/********** **********, and ******* ******* **** *** ******* reliable ****** *** *** ********* ** be ****** *******. *****, * ***** that ****** **** **** * **** large *********** ****** *** *** ******* to ****** * ******* **** ** more ******** *** ****** **** * traditional ***** ********. * ***'* ********* on *** **** "**** ******** *** useful" ********** **** ****** *******.

***** ****** ***'* ** * **** time. ** *** *********** ******** *** peak ** ** ** *** ********** wave, *** ***** ****** ** ***** of ****** ***** *** **** ******* may **** **** *** ******* ** ride ****** *** ***** **** *** refine *** ******* **** ********* **** is ****** ** "******* *********" *** less **** "****** ***** ***** ********". Those ** **** ***** ** ******* video ********* **** ***** ***** ** say **** * "****" ******** ********* is **** ******* ** **** ** detects *** **** ******* ** **** it ************ *******, ** ***** *** way **** ***** ***** ******* ** a ****** *********** **** ***** ******.

** *** ***** ** ** *****, it ** ********* (***) ** ********** evaluate *** ****** ****** ********** ******, like * ***** ********, **** ******** has * *** **-******** ******* ****** major ****** ** ******** ***********. ** is ***** * ********* ******* *******, and *** **** ****** *** *** feed ** ******* **** ********** *******, the **** ****** *** **** *** good *******. ** *** ***** **** this ***** **** ** **** ******** to **** ******* ****** *** **** not ** ****** ** ******* ** low-end (*****) ******* ** ****** ****** scenes *** ******* **** ********* ** quality ****** *** ***** * *****.

Agree: 5
Disagree: 2
Informative: 5
Unhelpful
Funny

***** ** ******* ********** ******** ***% accurate... ** ** ******* ** **** technology *** ** ******** ***% ********?

IPVM Image

Agree
Disagree
Informative: 1
Unhelpful
Funny: 2

* ****** **** ** *** ** the ***** ***** *******, *** **** everything * ***, *** *** **** 100% ********* ****. ** ** ******* above, *** ******** ** * **** system ** *** **********, *** ***-********** without **** ******** * ***** ****** of ***** *********.

***, ********* ********* ********* **** *** are ******* **** **** ** *** really ********** *** ******, ** **** specifically *** *****, *** ******* ********** up *** ******* ** ***** *********.

Agree: 5
Disagree
Informative
Unhelpful
Funny

* ** *** **** ** ****** is ***** ** **** ** **** this ************...

** ***** ******** ** ******* ******'* weapon ********* ******** *** *** *********** with *** ******* ** ***** ******* was ********** ** ** ***** ******* competency *** ******.

Agree: 3
Disagree
Informative: 1
Unhelpful
Funny

***, * *** ****** ** *** pivoting ** (** **** **) ******* detection *** *** **** ********* ** *** **** ******* **** ***:

IPVM Image

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

IPVM Image

Agree
Disagree
Informative: 1
Unhelpful
Funny: 3

**** ******** **** ** *** ******* looks **** * **** ******** *** seems **** ****** ** ***** **** work **** ***** ** *** ** that

Agree
Disagree
Informative: 1
Unhelpful
Funny

****** #* *** **** ***** *******! Historically ****** ******** ***** ******* ******* (alone) ** ****** **** (*** **** **** ***** **** ******) *** ** **** *****, ** does **** **** *** *** ******** is *********.

**** ** *****'* ******* (******** ****) demo **** * **** ******* ***** detector:

** ********, *** ******* **** **** just * ******* ******:

Agree
Disagree
Informative
Unhelpful
Funny

** ****** *** ******* *** ****** video, *******.

**** ********** * *** ********** *** **** ******* ***** ******** being **** ** ** *********** ****:

IPVM Image

Agree
Disagree
Informative
Unhelpful
Funny

*'* **** ** *** ** ****** and ********** ** ******* *** ** Evolv.

Agree: 2
Disagree
Informative
Unhelpful
Funny

****, ****** ** *** *********** ******* One *** ***** ******* ** **** segment, ****** ******** ******** ** ****.

Agree: 1
Disagree
Informative: 1
Unhelpful
Funny

*********, ******!

Agree
Disagree
Informative
Unhelpful
Funny

******: ******** ** **** *****'* *** and *** *** ***** ************ * ***** ********, ************ ** *** ****** *** **************** ** ** ** * ***** detector:

IPVM Image

**** ** *** **** ** ********* that * ***** ****** *** * company ** ******* ** *** **** is *** ******** ************ *******. ** ****** ******* ***** ** or **** *****, ****** ******** ****** Evolv ********** ** *******.

***, *** *** ** ******** ******* **** *************** "****************", **** ****** *** *** include * **** *** *** ******* recalled ******* **** ** *** ****** Act ******** *******.

Agree
Disagree
Informative: 1
Unhelpful
Funny
Login to read this IPVM report.
Why do I need to log in?
IPVM conducts reporting, tutorials and software funded by subscriber's payments enabling us to offer the most independent, accurate and in-depth information.
Loading Related Reports