Testing VideoIQ's Video Analytics (iCVR)

** **** ****, ** ******* *** of *** **** ********* ************* ** video ********* - ******* [**** ** longer *********]. ***** ****** ** ********* calibration *** ****-******** ********** *** ***** substantial ********. ***** **** ****** ******* offering, *** **** [**** ** ****** available], ** *** ** *** **** products ********* **** ********** * ******, DVR *** ********* **** *** ****.

Challenges **** *********

**** *** **** *** *****, ******** performance *** **** * ***** ******* topic **** **** ***** ******* ************* results.

********** *** ********* *** * **** of ****** *********** **** *******. **** huge ********* ********* *** ******* *** party ********* ******* *********, *********** **** worked *** *** ** ****** **** hard ********* ** ******.

Summary ** ******* ******** ****

***** *** *** ******** **** *** test:

• ***** ********* ** * ******* ** environments **** ********. * ***** ****** of ********** ********* ******** ***** *********.
• ***** ********* (******* ********) *** ********* very *** ****** *** * ****** of ********** **** ****** ** ********* considered.
• *********** *** ******** ******* *** ******** configuration; *******, ***** ****** ********* ****** limitations *** **** ***********.
• ******* *****-** *** ****** **-****** ******* temporary ********** ** ***********
• *********** *** ********** **** ********* ********** provided ** *******. *******, **** ** the **** ********** ********* ****** **** not ********** **** *** **** *******.

Test ***** *** ********

** ****** *** **** **** ** 80 ** **** ***** (***** ****** ***-*****) which *** ** **** ** ************* $1,800 ***.

*** **** ******* **** ******** **** software ******* *.*. **** ** ********* because *** **** ***** **** ** earlier ******* (*.*) ***** *** ********** less ******** **** *** ******* *******. Testers ** ***** ****** *******/*** *** newer *******.

******* ****** **** * ****** ** limitations ** *** ********** ******. ** did *** **** ** ****, *** or ****. ** *** **** ** a ******* ****** ** ********, *** not *****, ****.  ***** ** *** observations **** ****, ****** ** ***** and *******, ** ** *** ******* severe ******* **** ************* ******** ***** positives. ** ** ******* **** ******* can ******** ***** ********* ** *** ability ** ****** ******* ********* ** the ********** ** *** ***** ***.

Daytime ***********

** *** * ****** ** ***** in * **** **** ***** ** better ********** *** ***** ** *** scene *** *** *********** ****** *** day. ** *** ******* *****, ** were **** ** ******** ***** ******** at ****** ** ** *** **** wide (************* ** ******). **** ******** VideoIQ's ********* ********.

** ******, *** ***** ***** ** the *** ******* ** *** ***** for *****-**** ************ *** *** ******** of ********** ************ *********.

Nighttime ***********

**** ***** ****** **** ***** * few *** (*** ******* ***** * lux), ****** *********** ***** ********. ******** the ****** ******* ******* **** **** clearly ******* *** ********** ** ****** them. ** ***** ****** *********, ******** accuracy *** **** ********.

** ***** **** .* *** (******** in *** ****** *****), *** ***** became *********** ******** (***** **** ****** contrast). ** **** *********, ******* ***** be ****.

***** *** ***** ** ******** ** many ******** ************, ***** ****** ** careful ***** ********* ***** ******** ********** and ************* *** *********** ** *** built-in ****** *** ******** ***********. ***** VideoIQ ********** ****** ** ********** ************, this ** ***** *********** *** ** cost ** ************** ***********.
Narrow ***** ** **** ***********

** ****** ****** ** ****, ***** 35 **** ****, ************** ** ******* was ****. ** ** ****** *** moving ****** ** *********, *** ****** could ********. *******, ** *** ****** was ******* ******* ** * ****** pace, *** ****** ********** *** *** identify *******.

**** ** ** ********* ********** **** designers ****** ****. ***** ***** ********* ** more ***** **** *** ****** ******* areas, ** *** **** ******** ** a ***** ****, *******'* ******* *** be ****.

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

***** ****** **** ********. ** **** tests, ** *********** ** **** **** a ***** ***** ***** *** *****. The ***** ****** **** *** ***** seem ** ** ****** ** * few ****** ******:

• ********* *****: ********* ********* ***** *** detected ** * ****** (******* ******* were ***)
• ***** ********: **** ****** ******** **** periodically ******** ** * ****** (******* leaves **** ***)
• ********* *** ******: **** ***** **** we **-******* *** ******, * *** false ****** ***** ** ********* (*** system ***** ********* **** **** ** seconds ******* *** ***** ******)
• ******** *** ******: ** *** ****** was *******/******, ********* * *** ***** alerts ***** ** ********* (***** *** system ***** ********* ******* **********)
• ***** ***** *******: ********* **** ****** were ****** ** ** *** ** when *** ****** ******** **** *** to ***** **** (*** **** *****), a *** ***** ****** ***** ** triggered (***** *** ****** ********** ******)

*** ****** ****** ** ***** ****** will ********* ****** ** **** ******** conditions. **** *** *****, ** ***** expect ***** ****** *** ** ****** a ******* (*-*) *** ****** *** day. ** ***** *** ** ***** lower **** **** *** ***** ***** depend ** **** ******** **********.

*** ********** ************ * ****** ** these ******:

Object ********** ****** ***********

**** ******* *** ******** ** *******, they *** *********** **********. ** ************ can **** ****** ** ****** ********** and *** * ****** *** ******* that *** ******* ** ********** ** the ******** ***.

** *** *****, ****** ********** ****** worked **** **** *** ***** ******** had ******* ********, *** *** ****** of *** ******** ******. *** ********, searching *** * ******** ****** *** wears * ***** ****** *** ***** to **** ****** ***** *** ***** work **** *** ******** **** ********. However, ********* *** * ****** ********** who ******* ******* ***** *** *** go ** ********* ******* **** ********* camera ****** ** ******** ** ****** successful ******* **** ******** ****.

Analytics ************* *** ************ ******

***** *** **** *** ******** ************* options *** *******'* *********. *** **** one ** ******** **** *** *** options *** "******" *** "*******" ***** (see ********** *****).

*** **** ********* ******* ** ***** in ************* *** ************ ** *** presence ** ** ****-*********** ****** ******** the ****** ***** ****** ** (******** displayed ** * **** *** ********* on *** **** ***** ****). **** auto-calibrating ****** ***** ** ** *** cause ** *** ******** ***** ****** generated **** *** ****** *******. ****** that, ******* ******** **** *********** ******** over ****, ************ **** ***** ****** are ********* **** *** ****-*********** ****** is ********.

***** **** ****-*********** ******** *** *** 3-5 ****** ****** ****** *** *** camera, ** ** ********* *** ***** to ****** ****** ***** ** ******** and **** *** ******* ** ******** power ******* *** *** ****** *** eliminated. ** **** ****** ** ****, this *** *** ********* ** ** a *********** ****.

Contrast ** *******'* *********** ** ********* ******

***** **** ***** ******** ********* **** bullish ********* ******, * ** *******'* claims **** ***** ********* ********** [**** no ****** *********] ***** *** ** being ***** ** *** *******.

• ***** ** *******: ******* ****** **** they **** '********** ** ******** ************" where ********* **** ***** ****** *** the ****** ************* ******** *** *****. In *** ******** *******, *** ****** for **** ************* *** ********* *** it *** *** ****** ****** ***********. In *** ******* ******* (*.*), **** the ****** *** **** *******. **** could ** * ******** **** *** it ** *** ********* ********* *** stands ** ***** ******** ** ***** marketing ********* **** ** *** **** introduced.
• ****** ********: ******* ****** ****, **** their ********, ************* ***** ******** ******* to ******* ******** *** **** ******* searches **** ***********. *** ****** ** unlikely, *** ****** ****** **********. ** our **** ******* *****, ** * small *****, ****** ****** ******** ******** matches. *******, **** * ****** ***** (numerous *******) *** ********* ******/******** **********, such ******** ***** ******* ******** ** incorrect ******* *** *** *** ******* any ******* ******* (********** ** *** search *** **** ******** ****).
• ******** ******* ************: ******* ***** * number ** ****** ***** ******** ****** dynamic ***********, ******** ********** *** ****** behind *****/********. ***** ******* **** ******** targets ** *** ** ***** **********, it ********* ***** ***** ** ******* (sometimes *** * ***** ****** *** often *** * ****** ** *******) in ***** **********. **** ***** *** system *** ********* *********, **** ** not * *******. *******, ** *** are ***** *** ****** ** ****** loitering ** ******, **** *** ** a *******. *** ********* ******, *** bounding *** ***** ** ** ************ around *** ****** *** *** *** period ** ****. ** *** ******** box ********** **** *** * ******, then *** ***** ** *********. **** created ****** **** ********** ********* ******. With ****** ****** ** ******, *** system ********* ****** ******** ****** ******** (see *** ***** ***** ******* *****). This ***** ***** *** ****** *** for ***** *********. 

***** ***** ****** ** *** ****** core ********/*********, **** *** ********* ** establish ****** ************ ** ***********.

Recommendations ** *******'* ******** ************

** * **** *****, ********* ***** analytics ******* ** * ********:

• ********* ** ***** *********
• ********* ** ***** *********
• ***********/********** *** ******

***** *** *** ***** ** *********, I **** ***** **** ** *** report **** **** ** *******'* ***. As *** *** ***** ***:

• ***** ********* **** ******. ** ************ with * ********** ***** (**** ** remote **********) ***** ****** ****** **** without * ***** *******.
• ***** ********* ***** ** *********** ** certain **********. ** ************ **** *** to ***** *** ******* ***** **** (like *** ******** ** ***** ******** security ************) *** *** ** ********* with **** ** *** ***** ****. On *** ********, ** ************ ******* to ****** ******** ***** *** ****** the ******* ** ******** **** ****** but *** ***.