IronYun AI Vaidio v5 Tested

By Rob Kilpatrick, Published Aug 09, 2021, 11:56am EDT

IronYun has released their newest analytics, Vaidio v5, claiming improved accuracy and increased performance since our 2020 test, now including cell phone and PPE detection as well as vehicle make/model recognition.

IPVM Image

We tested these new generation Ironyun analytics examining:

  • How have person and vehicle detection changed?
  • Do animals and rain still cause frequent nuisance alerts?
  • Are cell phones reliably detected?
  • Can IronYun detect people wearing PPE reliably (hard hats, hi-viz vests)?
  • How accurate is vehicle make/model recognition?
  • How accurate is vehicle type (car, truck, van, etc.)?
  • Are animals accurately classified?
  • Has facial recognition improved in low light, IR, WDR scenes?

*******

*******'* ****** ********* ********* have ******** ************* ***** our **** ****, **** no **** ***** ****** detections ** ******* ** rain ******** ** ****** domes, **** *********** ****** in ***** *******. ************, person ********* *** ********** alerts ** ******** ******* of ******** ******* ** the ****, *** ********** offered (*.*., ***** **** when >* ****** ** present).

***** ******* *** ** longer ********** ** ******, all ******* **** ** our ******* **** ********** as **** **** **** confidence, ********* *****, ****, and ****, ***** ***** not ** ******* ** eliminated *** ******** ***********.

** ******** ** *** and ******** *********, ******* now ******* **** ****** and ***, ** ****, but ** ***** **** they **** ******** **** when **** ** *** hand **** **** *** ear. ************, ******* *** roughly ***-****** ****** ****** false *********. *********, *** detection ********* ******* ******** hair ** **** ** hard **** *** ****** t-shirts ** ****-********** ****.

******'* *** *** **** and ***** *********** ******** from ********* *** ******** issues, **** **** **** and ***** ********* ********* or ****** *******, **** with *** ******* ******* visible ** *** *****. Vehicle **** ********* (**** new) ********** ******** ****, trucks, *****, ***********, *** bicycles ** *** *****.

*******, ****** *********** *** improved ********** ***** *** original ****, *** *********** subjects **** ******* ********** drop ** ***, *** light, *** **** (**) scenes, **** ** ***** angles, ****** ****** **** masked ******** ****** ***** reduce ********** *** ****** false ************ ** *** tests.

Compared ** *********** ******/******* *********

***** ** *** ***** alert ********** *** ********* accuracy, *******'* *.* ********* performance ** ** *** with ***** ***-********** ********* from *** *****, **** as******************.

************, ******, ******* ****/*****, and *** ********* *** rare, *** ***** ** many *********** ********* (****** ****/***** *********** ******), ****** ******* ******** from ***** ****** ** all ** ***** *********.

Compared ** ******** ********** ****** *** ******** ****** ***********

***** ** *** *****, IronYun's ****** *********** *********** is ******* **********'* ********** *****************. *** ****** ******* angles, *** *****, *** IR *********** ********* ****.

************, *** **** **** to ********** ********* ******** wearing ***** ** ***************** **** ****** ********** drop, ****** *******'* ********** dropped **** ************* ** masked ********, ********* ** missed ************.

*******, ********** ****** **** rec ******* ** ********** lower **** *******, ** $500 *** **** *** channel ******** ** $***-*** USD, ***** ******** ***** as ******** **** *** ~$250-500 *** *** *******, with * *** ******* minimum ***** (~$**,*** **** cost) ** * $**,*** USD "******* ***."

*******

******* *** ** ********* street ***** ** ~$***-*** USD *** ******* *** live ********* (********* *********, facial ***********, ******** ******, etc.) *** ~$***-*** *** video ******, *** *******.

IronYun ********* *** ******

******* ******** ** **** their *** ********* *** appearance ****** ********* **** not ************* ******** ** version *.*, ** ** did *** ****** ***** analytics *** **** ******.

************, ******* *** ***** "** ************" ***** ******** **** rec *** *** ** compare * ********** ******'* face ** ** ** card (*.*., ******** ** or ******'* *******) ** confirm ***** ********. *******, templates *** **** ******** must ** ********* ** IronYun, ***** *** *** occur ** *** *+ weeks ** ******.

Animals ** ****** ******** ** ******

** *** **** *****, animals **** ********** ******** as ******, ********** **** approaching *** ******. **** has *** **** ********, with ** ******* ********** as ****** ** *** weeks ** *******.

IPVM Image

**** *** **** ** multiple ******, ********* *********** and ****** **** **** the ******, **** ******* were **** "*****"-******.

IPVM Image

All ******* ******** ** "****"

*******, *** ******* **** in *** ***** **** classified ** ****, ********* birds, ****, *** ****, and ********* **** **** confidence.

IPVM Image

******* **** ***** ****** will ** ********* ** future ****** *** *** not **** ****** ***** testing.

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

************, ***** **** *** droplets ********** ** *** camera's **** ** ****** triggered ***** ****** ********* as **** *** ** our ******** ****.

IPVM Image

Accurate ******** ****** ******

****** *** ****** ***** to ******* *** ****** of ******** **** **** enter * ***** ****** an ***** ** *********. In *** *****, **** worked **********, **** **** subjects ***********/********* ********.

** *** ******* *****, the ******** **** *** trigger ***** *** ***** person ****** *** ***** (highlighted ** ***).

IPVM Image

**** **** ** ****** this **** * *** of ~** ************ ********, with ****** *** ** trigger ** *.

Cell ***** ********* **********

******** ******* ***** **** phone ********* * ********* whether ** *** ***** held ** *** ****** at ** **** ** their ****.

IPVM Image

*******, **** ** *** held ** *** *** as ** * ****** was *******, *** **** phone *** *** ********.

IPVM Image

************, ******* * ***** box ***** ***** ***** positives, ****** ** *** at * ***** ********** and ***** ** ******** by ********** *** ********** of **** ********** ******.

IPVM Image

Bag/Backpack ********* ********

**** *** ********* **** properly ******** **** * person *** ******** ** wearing **** ** *** scene. **** ******** **** multiple **** **** ********** detected.

IPVM Image

PPE ********* ***** *********

**** ******* ******** **** PPE, **** **-********** ***** and **** **** **** accurately ********.

IPVM Image

*******, ********** ** *** tests, *** ******** ******** subjects ** ******* *** when **** *** ****.

IPVM Image

**** ********** ********** ** ******** settings, ***** ***** ********** with **** ********** (*.* to *.* *** ** 1.0 *** ** ***** above).

Improved ****** ***********

******'* ****** *********** *********** in ****** *** ****** is *********** *********, **** subjects ********** ************ **** at ***** ****** (**° angle ** *********/**° ********), with **** ****** ********** loss.

IPVM Image

Subjects ********** ** *** ***** ** ***** ********

*** ***** (~* ***) performance *** ******** ************* since *** **** ****, now **** ******** ********** by ~** ******, **** when ***** ******* ********* (~33°).

IPVM Image

Improved ** ***********

** *** ******* **** of *******, ******** **** frequently ******* ****** ** IR ******. *******, ** their *** **********, ******** in * **** ***** (~0.02lx) **** ********** **********, with * **-** ***** confidence ****, **** ** harsh ********* (~**°).

IPVM Image

Subjects ******** ******* **** *** ******* **** ***** ********** ****

******** ******* ********** ** hats ******* ******* *** scene **** ******** **** at ***** ****** **** only ~* ********** ****.

IPVM Image

**** **** **** **** at *** **** **** subjects **** ***** ********** with **** ~** ********** drop **** ** ***** downtilts ** ** *******.

IPVM Image

Subjects ********** **** ***** **** *** ********** ****

******** ******* ***** **** recognized **** ~**-** ********** drop. **** ********** ******** in ****** **** ***** the ******* ********* ** 70.

IPVM Image

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

****** ******* *********** *** problematic ***** ** ***** false *********** ****** **** setting ********* ***** **.

IPVM Image

Car ****/***** *********** ***** ************

*** **** *** ***** detection *** ********* ********, especially ** ******* ****** when *** *** ****** and ******** **** ******* visible.

IPVM Image

*******, *********** *** **** accurate **** ******** **** at ****** ******/***** *** or ** ********* ************ (e.g. ******** **** *** side) *** ~* ** every * **** **** missed.

IPVM Image

*******, ********* ****/***** *** simply *** ******** ** all, **** **** ******** were ******* *******.

IPVM Image

***** ********** **** **** IronYun, **** ********* * new ****/***** *********** ***** which ******* ****** *********, but *** *** ********* the *****.

Vehicle **** ********* ********

******* **** *** ********, this ********, ***, *****, bus, ****, *** ***********.

IPVM Image

Internal ********* *** ********

******* *** ******** ** option ** ****** *** events **********, ******** *** reliance ** * ******** VMS ** *** ** review ******** *****, ***** speeds ****** ********. *******, users **** **** ~* minutes ****** ******* *** event *** ** **** needed ** ******* *** footage.

IPVM Image

Versions ****

*** ********* ******** **** used ****** *******.

  • ******* ******: *.*.*-* / PowerModel-Pro-2.0

Comments (3)

**** **********, * **** with * *** **** that ***** ** ******* this *******.

******** * *** **** invited ** ** ********** to ******-**-******* *** *** *** function.** *** *** ** foraccess ******* to ****** * **** ****** **** ** ******** ****** *** ** ****** ** ******* *** ********* **** ** **** **** * ******* ***** *** ********** ** *******.

*** ***-**** ** * country **** **** ********* that **** ** ********* the *** ******** ** tandem **** * ******* management ******** **** ** customizable *** *** * data ***** ***** *** License ******. ******* **** be ******** ** **** with *** ******* ********** platform **recognize ********** ****** ******* ******. The existing camera sees the plate as it moves through the entrance, Ironyun pulls from the list of authorized users and triggers the gate to open. All can be viewed live of course. A **** *** **********!

*** ***** *** ** the ****, *** ***** in *** ***** ******* hears **audible ***** and it shows the vehicle in line on the screen who's license plate is not in the system. This gives the guard information about the plate and vehicle not on the list. The guard then verbally verifies what that guests business is on property, and then can add the license plate and notes into the system for whichever list they deem it should be associated with (can create specific lists such as: service worker, employee, resident, or not allowed on premises). The ***** *** *** *** ******** *** ***** ************.

**** * ***** ** recognized ** *******, *** list **** *** ***** is ********** **** - populates ** *** ******. Later *** *********** *** be ******** ** *report ********* and or exported by the Ironyun LPR portal and or be seen in a time line by clicking on the plate in Ironyun. The reporting feature will be used for accounting for this end-user.

*** ******* **** **integrated **** ****** *******, *** ***, ******* ********** ****** *** *** ***** ********** ******** that the staff uses to keep up to date information and accounting information. Then they only have to enter the information in one place.

** ****** **** * charm! *** ********** *** be ********** ** ****** be **** ** * standard *** ****** (**' from *** ***** *** level **** *** ******) though **** ** *********** (and ***** *********).

* ****** *** ****** used - ** **** up ****** ** ** are ******* *** ****** view ** *** *******.ALL ******* ******* ******, *** ****** * *** ******* ******* ** ******* *******, **** * ***** ****** ****, ******** ** * ****-** ** *****. So a typical LPR camera isn't as desirable. ***** ** *** ***-**** already *** * ****** hung ** *** *** side ** * ********* or ****, ** *** use **** ** ******* present *** *** *** function.

** **** ************* (******** camera ** *** ******** on *** **** ** the ******** ** ****) - ***** ******* ********** that ** ** ****** a ******* (*** **** associate *** **** *** model ****), *** **** it ********** *** *****. This ******* *** ****** that * ********* ******* could ***** ** ***** of *** ***** ** the ******, ** ****** the ***** **** ** authorized ******* ** ***** to **** *****.

************ ** *** *********** by *** ***-**** **** the **** *** *** -the '******* ********* ****' ***** ** ********* *** ***** ***** ***** ** **** **** **** ******* ** ********. No need to create that rule individually, it could be done by an associated list. If the landscaper was trying to sneak back on premises to use the pool, the gate wouldn't open. It is just as easy to remove a guest from a green list and move them to a red list, as in not allowed on property.

** ******* ***, ** key ****, ** ****, no **** *** ** isautomated *** ********* **** *** ******* ********. All instances can be monitored **** ** *** ******** *** ** ** * ***********, you see the instances on an associated (and included) ***, *** *** *** even ******* *** **** what ***** ****. ******** with * ******** ***** go ** *** ********** while ** ******** **** an ******** ***** **** an ***** ** ***** to *******.

* ****** ***** **** is ** ******* **********,the **** **** ** **** ** ******** ** ******** *** **** ** ****, ****** *** **** **** **** ******** - it really has that WOW ******.

*** ** **** ****** theID ************ for entry taking off (2 part verification with a drivers license or student ID etc. and facial recognition verification to gain entry into a gate or door. Typically deployed on existing video door stations).

** ** ***% ***** checking *** *** ********* for **** ********* *** a ********** **** **********. I ***'* **** ** sell ****, * ****** to **** **** ******* the **** ** ** amazing!

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

Agree
Disagree
Informative: 4
Unhelpful
Funny: 1

* ** **** ** independent ************'* *** *** Ironyun's ****** *** *+ years ***** ** *** Northeast **.

**** *** ** *** Vaidio ******* *.*.*.*-********** *** 2.0 ****** *******. * appreciate *** ****** ** comment **** ** **********.

***** **** **** ****** the ******* *** ** model? **** **** ** the ****** ********* *** LPR ****** ****** ** in *** *** **********. It ** **** **** in ** ** *** of *** ******* * casinos ** *** ******* after ********* ******* . I ******* ** **** point ** **** ** is **** ** **** while **** ** * cost ********* ***** *****. Works **** *** ****** 2 ** * ** IR ****** ** *** optimal ******** ** ******** position *** ** **** more ********** **** **** with **** *** ********* camera **********.

* **** **** ***** hand ** ******** ********* how * *** ** feature **** ********** *** then **** ******* ******** input, ****** ******, ******* improves *** ** ***** learns *** **** ******** and ************ ******** *********** over **** ** * future ****** *******. **** is *** ****** ** AI **** ******** *** all ** ***** ********* are ******* *** ***** in *** ***** ** ongoing **********, *****, *** after *** **** ******* via ***** **** *** customers. *** ********** ***** what * ** ******* about...it's **** ********* ******* vs ***** *** ******** after ********* ** * manufacturer's ******** ********.

**** **+ ** *********- Vaidio *** * ***** solution ******** **** ********** and ***** ****** *** grow **** ****** **** vertical ******* ** ******* a ***** **** **** new ************ *** ** forced ** ******* ** another ******** ********. ***-**** requirements ****** *** ** the ** **** ******** approach *** *** **** block ***** ** ********* with ******** *******. *** analytics *** *** ***** despite ********* ****** ...******* feels *** ***** ** processing *** ********** ** false ********* ** ***% greater **** **** ****** based ********* ***** ********** power ** ********** *******. The *********** ***** *** try ** **** ******* the **** *********** *** a **% ******** ** in *** **%+ ***** for ***** *********. *** get **** *** *** for. * ***** ********** analytic *** ******* **** more ******** ** ****** but ** ******* * question ** ********* ********** resources ** * ****** vs * ******.

***********: ****** ****** **** a ****** ****** ****** on * ******** ***** edge ****** ** ********* cameras ** ******* ***/** in * ******* ***** environment ***** * ****** container (******* ********* ****) running *** ** ******** analytics ** *** **** camera.

DIY ********: Not many suppliers offer "Do it Yourself" learning for AI Video analytics in the field. Please test the DIY application also. The DIY analytic allows the end user to customize training the system to detect a specialized alert. Examples we have experienced range from an airport who wanted to detect and alert from seeing debris like rocks or pebbles on a runway or in South Africa wanting to detect mountain lions. The US Park service uses Vaidio to detect bears or elk or bears and categorize them.

****** *** ******* ****.

Agree
Disagree
Informative
Unhelpful
Funny

****** ******** *** ***** post ******** ** ********** to ** ******* **'* mostly ********* *** *** primary ***** ** ******** on ***** ** *** 3rd ******* ** *** questions ** **** ********, not ****** *** **** money **** ***** ** these ********.

*******/**** *** **** **** before ******* ** **** emails * *** ***** ago **** *** *******'* partners. **** ** *** a **** *** ** egngage **** **.

Agree: 1
Disagree
Informative
Unhelpful
Funny
Read this IPVM report for free.

This article is part of IPVM's 7,216 reports and 962 tests and is only available to subscribers. To get a one-time preview of our work, enter your work email to access the full article.

Already a subscriber? Login here | Join now
Loading Related Reports