Avigilon Social Distancing Analytics Tested

** **** ******, ** share *** ******** ******:

• *** ******** *** ********'* distance ************?
• *** ** *** ******** between ****** *******?
• *** *** *** ****** monitored *** *******?
• *** **** ******** ** the ****** ****** ********?
• **** ******* ******* *** analytics?
• **** * *-****** ***** showing *** *** ***** it ***** *** **** not

Executive *******

********'* ****** ********** ********* performed **** ** ****** and ******* ************, ********* accurate ******** ************ *** alerts. ***** **** ** significant ****** ****** ********** violations. ** ********* ************ in **** *** ***-*****, and ** ********* ** to ***' (*****) **** the ******. ********* ****** detections *** ********* ******* identified ** ****** **** rare *** ********* ***** social ********** **********. ***** is ** ********** ****/****** for *** *********.

*** ******* ******** ** the ********* *** **** supported ************ *** *******, ***** **** ******** in ********* ****, ***** means **** ******** ******* do *** ******* ****. Additionally, **** **** ** a ****** ***** *** system **** ******** ****** where ****** ********** ** not * ******* (*.*. parents **** ********). ******** acknowledged **** ****** **** are ********* ****** ********** and ********** ************.

Overview *****

*** ********* *-****** ***** reviews *** *** **** findings *** ******** *********** issues:

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

********'* ****** ******** ************ worked **** ** *******, highlighting ****** **** **** not *********** * ******* minimal ********:

*** ******* ******* ******** distance **********, *** ** distances ** ** ***' (14PPF) **** *** ****** returned ********* ******** ************:

**** ****** ********** ***** returns * ******** ***** in ****** *** ****, which ** **** ****** testing ** ***** *** the ******** ** ********* distances **** *** ******:

Alert ** ********* *** **** **** * ******

******* *** ***** ** triggered ******* * ******, the ** ** ******* in **** ********* *** 2 ******** ****** ***** on *** ****** *****, even **** ***** *** multiple *********** ********* ******** limits:

** ******* ** ***** some ****** ****** * people **** ** *** closest ** **** *****:

******** **** **** **** working ** ********* *** UI *** ****** *** highlighted ******** *****:

[*** *******] ***** *** distance ******* *** ********** human ****** *** *** other ********** ***** ******* in *** *****. ****** are ****** ********* *** all ********** ****** ***** that *** ***** *** SD *********. *******, *** way *** ** *****, only *** ** ***** events ** ***** **** the ****** ******** *****. The ** **** *** working ** **** ** better ******* ************* ** multiple ******.

Self ******** *********** ********

****** ******* *******, **** the ******* ********* ** a *** ***********, *** distances ******** **** ************* longer **** ****** ********* between ****** (*.*. *' measured ** *.**'), ****** the ****** ***:

** ******** **** ** Avigilon *** **** ***** that *** ******* **** to ****-***** *** ***** to "********* *** ***** of *****":

*** ****** ******** ******** uses *** ****** ****-******** process ********** ** *** Avigilon ********* *** ******* the **** ***************. *** time ** ***** *** the ****** ** ******** self-learning **** ****** ** the ****** ** ******* that ***** *** ***** during *** ******** ******. Fewer ******* ****** *** period ***** * ****** self-learn ****. *** ****-******** period ** **** ** the ******** ****** ** calibrate *** ***** ** field *** ****** *********. While *** ****** **** start ********* ******* ***********, it **** *** ** optimized ***** ****-******** *********.

** ***** ****** ** hours *** ******** ********* significantly.

Measurement **** **** ****** ****** ** ******** ***

****** *******, ** **** multiple ***** ******* ** known ******* (*', *', 10') ** ***** *** camera ***********. ** ***** that *** ****** ***** not ******** *' ****** distance ****** ***** **** subjects **** ******** *** 6' ******:

******** ********* **** *** distances *** ******** **** the ******-****** ** *** bounding *****:

**** **** **** *** our *********, *** *********** is **** *** ****** of *** ****** ** the ******** ***.

*** ****** ******** ******* 6.1' *** *.*' ******* the **** ******** ***** separated ** *** *' object.

Dual ********** ***** ***** ******

** ***** **** **** or ********* ****** ********** would ****** ***** ******:

**** *** **** ****** around * **********/********* ***** the ******* *** ********* obscured. ** ******** ***-****** distance *******, **** ********** were ********* * ******** and *** ******* *********** impact.

Inanimate ******* ****** ***** ******

** ***** **** ********* objects (*******, *****) ******** as ****** **** ******** false ****** ********** ******:

***** ***** ********** **** rare **** * **** of ******* *** ***** create ******** ****** *** inflate ********* ****** **** long ******* ** ****.

ACS *** ********* *****

*** *** *** ********* "Covid *********" ****** **** in *******, ******** * quick **** ** * Covid-related *********:

• ******* *********
• ***** *** ****
• ****** ******* *****
• ****** ********** **********

*** ******* *** ** filtered ***** ** **** range *** ***-******* *****:

******** **** ***** ** "Entry ******* ******" ***** worked ************ ** **** if ********* ****** *** a ***-******* **** *** been ***, ***** ** Entry *** **** ****** detection:

*** ***** ******* ****** uses ******-***** ********* ********* to ***** ********* ****** an ****, *** ****** Distancing *********. **** ***** if ***** *** ** distance **********, *** * government *** *** ******* permitted ****** ***** *** indoor **********, **** *** be ********* *************.

**** *** ****** ** a **** ** *******, the ********* ********* ****** well, ******* ** **** entries *** ***** ** well *** ****** *********.

** ********** ************** **** **** ************** ** **** ******** in *** ***** *********,***** ********* ***** **** in ********* *** ********* where **** ***** ** used (******* **** **** lighting ** *** ***** and ****** *** ******). However, **** ***** ** (for **** *****) ** in ***** ***, **** detection ****** ******.

Version/Cameras ******

• ******** *** *.* *** 7.10
• ******** *.**-***-***-** ******** *.*.*.**
• ******** *.**-***-***-** ******** *.*.*.**