ZeroEyes Gun Detection Startup

By: Sean Patton, Published on Jul 16, 2019

A gun detection video analytics startup, ZeroEyes, is being led by a group of 6 former Navy SEALs, aiming to "save lives" by using AI to assist first responders of active shooting incidents.

** ***** ** ********, to ********** ***** *********, go-to-market *****, *** *** they ******** ***** ********* in ****** ************.

** **** **** ** look **:

*** ** ********
**** **** ** *** Detection
**** ********* ** ******** Targeting
*** **** **** ** Cost
*** **** ******** **** Analytics
**** ******** ********* ********

ZeroEyes ********

******* ** ****, ******** is ************* ** *********** and ******* ** ****-**** employees, ********* * **-********. ZeroEyes ** **** ***** **** ******, *** ***** ** years ** * **** SEAL ***** ** ******* his *** ***** ******* ******. *** ******** ** their********** ******** **** ** ***** military **********, **** **** listing ******** ***** ** combat *******.

******** **** **** **** initially ************ *** **** be ******* * ****** fundraising ***** ****. **** plan ** ****** *** funding ** ********* ********* development ** ****** **** scientists/developers, ******** ***** ******* partner ******* *** ********* marketing. **** *** ********** as ******** ******** ******* ***** Small ********.

Real-Time *********** *** ****** ******* ********

******** **** ***** ******* aim ** ** **** an ****** ******** ******** before ** ******. ***** focus *** ********* **** on *******, ******* ** the********** ****** ** ** school ********* ***** ****, *** ****** ******** concerns *** ******** **** member ******* ******* ** schools.

**** **** **** ***** combat ********** ***** **** a ****** *********** ** active ******* *********, *** an ************* ** *** "scary *** *******" **** are, ******** *** ** a **** ** *********** for ***** ********** (*.*. How **** ********, ***** are **** *******, **** type ** ****).

**** *-****** ********* ***** pitches ***** ********:

Gun ********* ******* ** ****** ******* ****** *** ****-********* *******

******** ****-**** *** ********* platform ** **************, ***** ** * deep ******** ********* ****** based ** ******** ******** of ****, *****, *** Faster *-***. **** ****** and ******** **** ***** an *** ****'* ******** IP ****** ******. *********** resolution ** ***** (** higher), *** **** ******* 7 - *****, ********* a ******* ** **% exposure ** * *** for *********.

**** ******** ******* *** rifles, *** **** **** they ********** *** **** for **** ****** *** types *** ******** ** active *********, ****** **** "95% ** **** *** black ******* ****** ** semi-automatic *******". **** **** they **** ****** ****** at ~****/*** *** ******* at ~****/***. **** ***** their ****** *** **** effective ** ********* ********** (chrome) ** "*****-*******" ****.

Custom ***** ******* ********

******** **** **** **** trained **** **** *** models, *** ***** *********** resources ******** *** ******** their *** ****** ******** of *** ****** ** ~10,000 ******. **** **** this *** **** ********* for *** ***** ******** as ******** ********* ******** for **** ** *** low ******* *** ******* in ********.

**** ***** ********* ***** own ******* ***** **** is *** ** *****, because ** ***** *********' personal *** *********/******* **** handling ****.

False ******** ********* ********

*** ** *** ******* model ******** ***** ***** by ******** ** *** reduction ** ***** *********. They ******** ** ******* false ******** **** ** 2 *** ***** *** 25 *******. **** **** that ******* **** *** using ******** ******* *** primary ********* ** *********/******* lighting. ** **** ** optimization, **** **** **** document *** *** **** a *** ** ********, time ** *********, *** what **** ** *** was ********.

************, **** **** **** have ******* *** ********* to ***** ***** ********* with ******** ******* *** objects (*.*. ******** ******, lacrosse ******, ****** *******).

Primarily **-******* ********* / ***** ******

******** **** **** ********* sell ** **-******* ********* with *** ******** ****** hardware *** ********** ***** analytics. **** **** **** are ********* ** ******* different ****** ****** **** RTX2080ti ** ***** *****. They ** *** ***** a ********-**** ******.

**** **** ***** * cloud-only ******** *** ***** it *** **** **** deployed ** **** ************. They **** ***** ***** system ******** "***** **** forwarded *** ****** ** complicated ****-**-**** ***** *** tunnels", ********* ** *** end ****'* ******* ******** policy.

**** ********* *** **** to *.* **** *** camera ********* ** ****** density, ******** **********, *** camera ***********. **** **** the ***** ******** ******** the **** *************** ** 1080P *** * - 15fps ** **-******* *******.

Police ********** ***********

******** **** * *** operational ******* ** ***** system ** ************* *** involvement **** ***** *** enforcement. **** ******* * web-based ****** *** ****** SMS ***** ************ ** ensure **** ***** ********** are ********* *********** ******** from *** ******** ******, not ****** ** **** on ****** ********* ** systems ** ******* **** is *********.

**** **** **** **** have * ****** ****** monitoring ****** ** ***** roadmap, ** ****** ****** departments *** ****** *** school ** **** **** contact **** *** *******.

VMS *********** *********

******** **** **** ****** RTSP ******* *** ******* alarms/events **** *******, *********, Exacqvision, *****, *** *****. They ** *** ******* bounding ***** ******** ** configuration ** ******** ** the ***'* *********. ******** web ****** ** **** for ************* *** ******** video ******** *** ***************.

**** ***** **** *** working ** ****** ************ for ********, ***** *** Panasonic/Video *******. **** *** not ***** ** *** VMS ************'* ******** ** a ********* ** ********* integration.

*******

******** ** * **** subscription **** *** ****** per ***** *** $** - $**, *** * customized ********* **** **** will **** ** ******* significantly ***** ** *** NVIDIA ******** ********* *** the ****** ** *******. They ***** **** ******* pricing ********* *** *********, and ******* *** ****** decreases **** ****** ******** lengths.

******** **** **** *** currently ******* **** ***** school *********, **** * minimum ** ** *******.

Channel *********

******** **** ******* **** are *** ** *** market **** *** ********* marketing ****** ** ***** end-users, *** *** ******* with * *** ********-***** integrators ** *******, ******* and ************ **. **** noted **** **** *********** do *** **** *** niche ********* ** **** and ******** *** ********* video *********, *** ***** larger ***-***** **** ******* that ******** **** **** their ********** *** ****** adjustments *** *********.

*******

******** *** **** ****** expertise ** ****** ******** scenarios *** ********. *******, their ******* ************ ** limited ** ******** ****** or ******* ******** *********** experience. ************, ******* **** limited *******, *** ******** is ******* ** ********* add-on ******, ***** *******, licensing, ******* *** ************** are **** ***. **** has ********* **** * difficult ***** ***** ** scale. ******** **** **** in ******** ** *******, they **** ******* ***** deployments ** ***** ********* (e.g. *********, **** *******, public ******).

Comments (27)

Undisclosed #1

07/16/19 03:11pm

** **** ***** ** be ******* **** ******** Dual **** ***** *** a ***** ** ***** V100s ****** *** ** a **** *** ***** so ** *** *** some ********* ******* *** benchmarks *******?

***** ** ** ** expensive **** ******* *** can **** **** ***** systems *** ****? ********* and ***** ****** ** IPVM *********?

Ethan Ace

IPVM | In reply to Undisclosed #1 | 07/16/19 04:54pm

** ******** **** * new ****** **** * Tesla ** ***** ** recently *********, ****** ** definitely ** *** * 256GB *** ***** (***...* think ** *** ****** slots *** ****). **'** currently ******* * ****** other ****** ****** ** get **** ** ******* it ** *** ** will ** ***** ** for **** **** ******** tests **** ****.

server bench setup - with ethan Video

Undisclosed #1

In reply to Ethan Ace | 07/16/19 08:40pm

*******! *** *****!

Slava H

07/16/19 03:34pm

**** **** ** * gun ***** ** ** on *********** ***** *** positive *********?

*** ********** **** *********** works ** **** ********** small ******** ******-*.*** @ 1080p, *********** **** *** is ******* ***** *** fast ****** ******.

Sean Patton

IPVM | In reply to Slava H | 07/17/19 01:29pm

******** **** ** **** as ** ***** **% of *** *** ** visible, **** *** ****** rifles ~****/*** *** ******* at ~****/*** **** *** camera.

Slava H

In reply to Sean Patton | 07/17/19 07:23pm

** ***** ** ****" **** **** *** cloud ******** ******** *** same *************** ** ***** and * - ***** as **-******* *******." *** this "

***'* **** "*******" *** camera. ** ***** **** on *** - ******. 20cm *** ***** ** 36x23 ******. **% ** it - ***** ******. **** is ***** ****** - no ***********, ***** *****, etc..

****** **** ** *** in ****** ******** *** proper ****-**** ********* ******** would ** **** *******.

Timothy Sulzer

In reply to Slava H | 07/20/19 04:12pm

********** *** ********!

* ***** ********** - PPF/PPM ******* *** ****** dependent ** ***** ******* like ***********, ******** ****** and ***** ******* ***** are **** ** ********.

*** *** ************, *** majority ** ******* **’** worked **** ** *** have ****** **** ~** degrees. *** ******* **** shows ** **** ******* 17PPF/55PPM **** * ***** area ** ~***-**** ** make * *********, *** we *** ****** ** further ********* **** ***** FOV ******* ~** *******.

*** ******* ******* ** the ****** ****** ** detect ** *********** - reducing *** ******* ******* of * ******* ****** at **’(***) *** * larger ****** **** *** same ********* ** * larger ****** ** **’(**)

** *** **** *** pixel **** ** *** detection ** ** * more ********* ****** **** PPF/PPM?

**’** ***** ***** ******* testing **** **** - what ***** ******* ** you **** ******** ** test?

Slava H

In reply to Timothy Sulzer | 07/23/19 09:29pm

***** *** *** ***** Timothy.

*** ******* ** *** detection ** *** *** and **** ***** ** "regular" ***** ***** ******** camera ******* ** **** of *********** ** **** good *********.

*** ******* *** *** saying:"*** ******* **** ***** we **** ******* *****/***** with * ***** **** of ~***-**** ** **** a *********". **** **** ******* **** gun ***** ** ** sqrt(500px)=22.4px ** ****(******)=*****. **** would ** ******* ******* 1/2 *** * ****** in **** *** *****. Unless * **** **** numbers *******. ******* ** 25cm ****** ** ***** would ** ******* ******** (55/4=13.75). *** ******** *** are ***** **** **, image ***** ** ********* and ** **** ***** even **** **** ****.

** * ** ******* what **** ** *** gun(handgun) ** ****** ****** Yolo ******* *** **** created, ******** ** **** correct *********.

Undisclosed End User #2

07/16/19 06:29pm

***********, ** ****** ***** of, ** **** *** intel ** ********'* *** detection *********? * *** just ***** * ************ from ** ******** *** recently, ********** *********** *** same ********* (****** ********* of * ***, ********** an *****/***** ** * VMS), **** **** *** developing, ** ** ********* sometime ****** *** **** year ** *.

**** **** * ******, their ******* **** **** account *** **** *** shape/size ** *** ********* weapon, *** **** ****** on **** ****, *** a ****** *** ******* the ******. ********** ***** version **** ** *** camera ****, **** ********* you ** ******** ***** equipment *** **** *** any ******* *** *** already ****.

** *** **** **** had *** ******* ** run ** ** ********* (assuming *** *** ***** VMS), *** * ***'* recall, * **** ******* off ** ***** ** the ************ ***** ** would *** ** **** us ** ***** *** platform (**'* *******... ** get **). ***** **** guy ******.

Undisclosed #1

In reply to Undisclosed End User #2 | 07/16/19 08:42pm

* ***** **** ** see ** ******* ** umbrella **** * ***** could ******* **** ******** rule.

Undisclosed Integrator #4

In reply to Undisclosed #1 | 07/17/19 12:30pm

*'** ******** **** * demo ** **, * have *** ********* ** though. *** ****, **** held ********* *** *********** like ** *** * gun, *** *** ********* didn't ****. *****, ** was * **** ***********, not **** *** ** was **********.

Undisclosed #3

07/17/19 05:45am

*** ****-***** ************* ** using **** **** ** object-recognition ******** **************** *** *******/********** **** in ***** *** ***** above ** *** :** mark:

"That ******* ** ***** ** **** *** * ******..."

** ********************* ** ****** ******* scenarios? **** ********** ******* the ******* ***** **** get ******?

* ** *** ******* that *** ****** *********** technology ***** **** ** this ******* ***** ** described - *** **** existing ************ ******* *** be **** ** *********** with **** ******** ** recognize ****** ********....

but *** **** ** * ********** *** **** ******** ******** ******* ** **** ***** ****** ******* **********?

** ***** **** ****** much ****** **** ******* (and * ****** **** story) *** *** ******** available ********** ** **** their *** *** *********** analytic *** ***** ***** own ****** ******* ******** to ******... *********** ** 'save *****'.

*** ******* ** *** all ** *** ******** industry *** **** **** very **** ****, *** we ******* ************* '*********' **** *** seemingly ******** ** ***** exactly **** ** *** all ****** ** ** day ** *** *** out... **** **** ** empirical ******** **** *** 'solution' ***** **** ******************* ** ********** **** the ******** *** *********** designed ** *******.

Michael Lahiff

In reply to Undisclosed #3 | 07/20/19 06:25pm

"That ******* ** ***** ** **** *** * ******..."

"** **** ****** *********** ** ****** shooter *********? **** ********** conceal *** ******* ***** they *** ******?"

***, ** ** *********** for ****** ******** ** take *** * ****** before **** ***** ** people. ** *** **** of ***** ****, ******** Beach, *********, ******** ****, and ******** ** **** as **** ******, *** shooter ******* *** ****** long ****** ***** **** fired. ** **** ***** the ******* ****** ** with ****** ******* **** parking ***.

"but *** **** ** * ********** *** **** ******** ******** ******* ** **** ***** ****** ******* **********?"

*** ********** ******* ***** sight *** ***** **** can ** *******. ***** is ** *********** ** prevent ** ****** ***** from ********* ** ** SRO/police ** ***** ***** in * ****** ******. (picture ** *********, ******, location, ****) **** *********** is ******** **** *** are ***** **** * chaotic & ******* *********** like ** ****** ******* scenario. **** ******* **** force ***********. ******** **** to *** ******* **** to *******, **** ********, is ********.

Undisclosed #3

In reply to Michael Lahiff | 07/20/19 07:15pm

"The ********** ******* ***** ***** *** ***** **** *** ** *******. There is an opportunity to prevent an entire event from occurring if an SRO/police is given intel in a timely manner. "

*** ****** ******** ***** contains *** **** **** qualifier '*** **' - in ***** ** *** up **** ********* **** 'there ** ** *********** ** *******' - which i have far less confidence in than you do.

*** **** **** ******* first ***** *** ***** shot ***** **** ****** ** be ******** ** ****** digit *******.... ********* *** ******** useless ******* *** ***/********** ******* *** ***** ** a '****** ****** ******' to ** ******** ** 'prevent'.

* ***** *** ********* analytics *** *** **** valuable ** *********** ***** stick ** ********** - as *** ****** ** displayed (*** **** ***** not **** *****) ** order ** *** *** clerk ** **** *** register.

*** ** ***** **** that **** **** **** spend*** **** ******* '*********'************ ******* ****** ******** than ********* ********* **** spend ** ********* ** prevent *********** ***** ***** ups.

Undisclosed #3

In reply to Undisclosed #3 | 09/06/19 07:20pm

** ****** ***** ** this *********** **** - though *** ********* ** pretty **** ******* *** same - ** **** all ***.

************

**...***** ******** ** '******'?

*** ******** **** *** team ** ****** ***** or *** **** **** no ****** *****?

***** **** ** ** audio ********* ****** ** visible *********, *** ********* is ***** *** ****: "We *** **** *******".

** ******** ******** ** claims **** ** ***** is ***** ** *** observation **** ***** ********** marketing ******* **** ** focus ******** ** ******** vs *** ****** *********** analysis.

Undisclosed #3

In reply to Undisclosed #3 | 09/07/19 03:16am

**** ** *** ***** time **** * **** ever ******* ******* *** voting *** ** ** comments '*****'.

*** ******* ** ******** solutions **** **** ** actual ***** ** *** funny ** *** ***.

*******, * ***** **** is **** *** ** many ******** ** *** humans *** **** *** can ****** ** ********* to *** *** '*********' that ****** ** ** tailored ** ********* ******** vs ******* *** ******** problem ** ***.

Undisclosed

07/18/19 03:08pm

* ***'* **** ** you **** ** *** or * ******** **********, if ***'** * ******** vendor * ****** *** to **** *** ***** to *** *** "***'* Encrypt" ************ ** **** website. ****'* **** ****** used **** ****** **** armor ** *-***. *'* sorry *** ** **** cloud ***** ** ***** call *** * *** to ****** ** ** have * ***** ******** internet ********.

John Honovich

IPVM | In reply to Undisclosed | 07/18/19 03:14pm

** ***'** * ******** vendor * ****** *** to **** *** ***** to *** *** "***'* Encrypt" ************ ** **** website

*** *******:

***:

** *** ** ** certificate **** ********. **** said, **** ** *** concern ** * ******* using * **** *********** for ***** *********** ******** website?

Undisclosed

In reply to John Honovich | 07/18/19 03:24pm

** ***** **** **** information ********** *******. ** demonstrates ** *************** **** of ******* *** *** concept ** ******** **********. It ******** *** *** team ****** ***** ******* has *** ****** **** (yeah * **** **** low ****** **** ***** a **** ****, *'* not ***** **** **** either.) ** ******** **** don't **** ** ***** whole ******** ******** ** a ***** ******. ** suggests **** ** ****** the *** *** ***** you ****** ** * summer ****** ** *** marketing ********** ** ****** a *** **** ** five ******* ** **** dollars ****** *** ** you ***'* **** ** dollars ** *** * real ****. "****'* *** us ****'* *** ********* department" **** *** *** it *** ****.

*** * **** ****** this ******* *** ** we **** ****** ***** down *** **** *** vendor * ****** ** get * **** ** edgewise ;-)

Michael Lahiff

In reply to Undisclosed | 07/20/19 06:42pm

******- ********** *** ******** to *** *********.

*** *** **** *** not ******* *** ****- they **** **** ******* on *** *******, **, & ********.

**** ** ** **- no *** **** ******.

** **** ******** ***** to ***** ********. ** team *** * **** to ****** ***** ******* on *** **** ***** fighting- ******** ********, **********, accuracy, ***** ********, ***.

*** **** **** **** be ******* ** *** time.

******,

****

Undisclosed

In reply to Michael Lahiff | 07/21/19 04:18pm

***** *** *** **** response. *'* **** *** have **** ** **** on. *** *** *****-*** in **** ***** **** us ****. * ***** say **** "******* *** market's *********" *** ** your **-** **** *******.

Undisclosed Manufacturer #6

In reply to Undisclosed | 07/19/19 10:49am

***'* ******* *** *** of *** **** *********** authorities ** *** *****. If *** *** *** domain, *** *** ***** it, **** **** **** you * ***********.

Undisclosed End User #5

07/18/19 06:21pm

** ****** ** **** ******** ** these ***** ** ********** active ******* ********* ******* and * ******* **** deserve *** **** ********* ** **** alarm systems. Without **** *********, **** are **** ******* **** that******* ** ****** *** *****. Because **** ****** ****** ** visibility ** * *******, I ***** ** **** **** be ***% ******** *** will **** ****** *** another tool ** *** ****** of ********, ** ****.

**** *-** ****** *** **** at **** ** *** of ***** "****** **** *******" solutions, *** * ***'* see *** **** ****** mitigates *** ***** ** someone ******** ** ***, but I ***'* **** *** unique *********** ** * combat SEAL.

* ******'* ***** ****** funds ** **** ****. We ** *** ******* **** built-in ******* ********* *********, because ** ** ******** with *** *******. **** another **** **** ** ****** never rely **.

**** *** *** **** on ****** ** ****** shooter *****? ****** *** good ** ********* ********, especially ******** ******** **** an ******* *******. ****** will **** *** ** a ***** *** *** call ****** **** *** very good *** ********** *********** for ***** **********. ***** and ***** **** ****** ** quickly **** **** ****** when ******* *******.

Undisclosed Integrator #7

07/21/19 02:40am

**** *** **** **** is ********** ** ****** knowledge, ****** ******** **** just **** ** ******* their ******* ***** **'* time *** *** ****** to *** *** *****. Shouldn't **** ** ******** to *** *** **** what's ***** ** ** that *****.

* ***** **** **** should ** ******* ** an ********* ****** **** as * *****-**** ********* measure.

Brandon Knutson

IPVMU Certified | In reply to Undisclosed Integrator #7 | 07/22/19 02:52pm

***** **** ******* **** ******* their ******* **** **** to ******** *** ****** *****. **'* predictable **** **** **** ****** conceal ***** ******** ***** **** for *** *******, **** or ******* ********.

Undisclosed Integrator #8

07/24/19 12:57pm

******** ***** **** ******* would ****** ** *** way "******" ****** ****** when ********** **** **** shootings **** * ******** of **** ******* ******** take ***** ****** ***, sometimes ** *** ******* lots ** ******* **** before **** *** ** the ****. ***** * agree ********** **** ******* will **** ******* *********, for *** ****** * few ******** ******* **** ZeroEyes ******** * **** chance ** *** ***** of * ******** ******* of ********** *** ****** during ***.

Undisclosed #3

In reply to Undisclosed Integrator #8 | 07/24/19 01:17pm

"********** **** **** ********* show * ******** ** mass ******* ******** **** their ****** ***"

**** **********? ****** ******* evidence.

Read this IPVM report for free.

This article is part of IPVM's 6,534 reports, 880 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.