SAFR Facial Recognition Inside Axis Camera Tested

By Rob Kilpatrick, Published May 10, 2021, 10:35am EDT

RealNetworks' SAFR entered the security market giving away their analytics to schools for free. But how does their face rec compare to enterprise-focused competitors like Anyvision or BriefCam?

IPVM Image

We tested SAFR Inside on Axis' deep learning Q1615 Mk III, and in this 20-page report, we examine:

  • How does it perform in real-world surveillance scenarios?
  • Does it falsely identify a person as a different person?
  • Does it miss faces walking through the scene?
  • Can it still recognize the faces of people wearing sunglasses, hats, hoodies?
  • How does it perform at night in low light (~2 lux) or dark/IR (~0.02 lux)?
  • Does it detect anything other than human faces?
  • Is the user interface easy to navigate?

*******

** **** * **** of **** *******, **** face *********** *** ******** in *******/**** ***** (~***+ lux) ******, ********* ***** horizontal ****** (** ** 90°) *** *********** ******** (30-45°), **** ** ***** positives ** ****** *********** of ********** ********.

*******, ** *** ***** and ****/** ********* ******** degraded *************, ********** **** the ******* *** ** even ****** ****** ******* the *******. ************, **********, hats, *** ***** ******* confidence ****** *** ***** of ********* *** *******. Wearing ******** ** ***** items ************ ******** ** missed ************, ******* ********** users ** "*********."

**** ******** ~***-*** *** for ***********, ************* ****** than ********* **** ** Anyvision ** ********, ***** required **-***. **** *** 1080p **** ******, **** worked *** ** * maximum *** ***** ** 12 ** ** **** / ~* ******.

************, ****'* **** ********* is ********** *** ************ detects ********* ******* ** faces, **** ** ********** or ******, ********* ** numerous "****" ******* ** its ***** ****, ****** it **** *** ******* to ********* ***** *******.

*********** ********* **** ** gender ** *** **** inaccurate, **** **** ******** shown ** ****** *** vice *****, ******* ******** on *********, *** ******** ages ********** *-** ***** higher **** ********' ****** age. ********** ********* **** identified ****** **** * neutral **** ** ******* accurately, ****** ****** **** a ******* ********** **** the ******** ** ****** identified.

*******, **** ****** ************* lists ******** ******* **** text *** ********* ****** (e.g., "**********.********-*********: *.*") ******* of *** ********* ********* used *** ****'* ******-***** settings, ****** ** ********* to ****** ******** ******* experience ** ********** **** SAFR.

Relatively ***** ****** ****

** *** *****, *** SAFR ****** ****** ******** 100-200 ****** ******** **** the ****** ** **+ days ** *******, ********* 5 **** ********* ** the ******, **+ ********* of ***** ********** **** our *******, *** ***+ visitors ** ****** **********. All ***** ******** **** added ** ****'* *********, along **** ***** ********* of **** *** ****** photos ** **** ********.

*******, ***** ********** *** relatively ***** ******** ** many ************ **** ** likely ** ** ***********, such ** ******* **** several ******* ********, **** staff *** *******. ******* in ****** *********** *** vary, **** ***** *********** issues *********** **********.

Cloud ********** ** **-******* ****** ********

**** ****** ** *** a ****-********* ****** ***********. Instead, ** ******** **** capture *** ******* ********** which ** **** *********** to * **** ******, either **-******* ** ** the ***** *** ***********. Because ** ****, ** an **-******* ****** ** not *******, *** ****** must ** ********* ** the ******** ** ** detection, *** ***********, **** be *********.

Pricing *** ************

**** ** ******** *** camera ******, **** ******* different *** **-**** *** cloud ********* ******* *** available ** ****** ************ or ***-**** ********.

  • **-**** ******** *** ** purchased ** * ************ ($69 ***/***** *** ****** MSRP) ** * ********* license ($*,***/******).
  • ***** ********* ******* *** available ** ************ ****, priced ** **** $*,*** USD *** ****** ($***/***** per ******).

**** ****** ** ********** with **** *******, ********* a **** ******** ********** Unit, ********* ******* ** the***** ** ***, *****-** Mk ***, ********-***.

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

** ***** ** ****** recognition ********, **** ************/***************** ********* ** *** tests, **** **** *** or ** ***** ************ (one ******* ******** *** another). ********* *** ********** more ******** **** ******** wore *******, *****, ** other ***********, ****** ******** performed ****** ** *** and *** ***** ******.

*******, **** **** **** does *** ******* *** video ******** ** ******* object ******* ** ********, such ** ********, *****, direction, ***, ***** **** BriefCam **** ********* ****** for ******** ******.

Few ******** ** **** *** / ******** ****** *********** ********

******** *** **** ****** app **** ** *** Axis ******, * ******** application ** ******** *** configuration, ********* **********, *** monitoring.

IPVM Image

** *** **** ****** app ******, *** **** options *** ** ******* overlays (******** *****) *** identities ** ********** ******** or ***.

IPVM Image

Configuration **********

** ***** ** ****** recognition ******** *** *** SAFR ****** ***, ***** must ******** ******* ******* levels ** *******-***** ********, e.g., **********.********-*********, **********.******-******, ***.

IPVM Image

***** ******** *** ********** into * **** ******* GUI **** ***** ****'* server-based ********* *** ******** instead ** *** **** app. *** *******, *** same ******* ***** ***** is ******* ******* **** a ****** ****** *** adjustment ** ****** ********:

IPVM Image

***** **** *** ** easier ** ***, *********** is ***** ***** *** inexperienced ***** *** ****** introduce ***** ******.

Simple ***** **** ******

****** ****** ** *** watch **** ** ****** in *** **** ******* app, ** **** ******** the **** ** *** a **** *** *** people *** *** *** user *** **** **** add ***** ** *****+******** face ********.

**** **** ***** **** input *** ******'* **** manually, ** **** **** not *** *** **** name ** *** ******'* name.

**** ******** * ********* image ******* ******* **** will ****** *** **** of * **** ********* image, ***** *****.

IPVM Image

** ********, **** ****** list *** *** ***** duplicates ** * ****** to ** *****, ******* a ******* ** *** user ***** ** *** the **** ****** *****.

IPVM Image

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

** * **** ** testing, **** ******** ** false ********* ** ***, night, *** ** ****** using * ******* ** over *** ****** *** 100-200 ****** ******* ******* the ***** **** ***** SAFR ******* ********.

IPVM Image

*******, **** ******** ******* intended ** ******* ***** issues ********* *** ********** of ***** ********, ********* below.

Solid **** ***** ***********

** * **** ***** scenario (~*****) **** *** solid ***********, ********* *** recognizing ****** ******* ******** at *** ****** ** at * **° ***** even ** ****** ******** heights ** ~**' **** harsh ******** ** ~**°.

IPVM Image

No ***** ********** ** **°

****** ******* **** *** camera ** **° **** not **********, **** ** lower ******** ******* ** ~7' *** ******** ** ~16°.

IPVM Image

Reduced *********** ** *** *****/****

*********** *********** *** ******* in *** ***** (~***), people **** **** ********** at ***** ******** ******* of ~*' **** ~**° of ********, **** *********** straight ** *** ******. Even ** ***** ******** heights ****** **** ********* still ********** "********".

IPVM Image

************ ** * **** setting (~*.****) **** ** performance *** ******* ** low *****, **** ******** recognition ** ***** ******** heights ** ~*' **** ~19° ** ******** *** approaching ******** ** *** camera. ************ **** ***** not **********, ********* ********* a ****** ** *** person **** ** * stranger.

IPVM Image

**** ********* ********* *** proximity ********* ********* ******* to .** ** ***** for ***** ********** *******. This ******** *********** ** harsher ****** ** *** light, *** ********** *** higher ********** ***** *********.

IPVM Image

**** **** **** ******** on *** *****/** ***********:

**** *** *** ******** an ** ******** ***** for ********. *** ******** benchmarks *********** * *********** degree ** *********** / stability ** ***** ******.

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

**** ****** **** *** include ******** *********. ** our *******, ******* **** the ****** **** * printout ** * *******'* reference *****, *** *******'* photo **** **** * person's **** *** ********** as *** *******.

IPVM Image

Reduced *********** **** *******/****

*********** *********** *** ******* when ****** **** *******, with *********** **** ** shallow ******. **** ** shallow ******, ******** **** occasionally *** **********, ***** as ********* ** ****.

IPVM Image

**** ********* ******** *** proximity ********* ********* * .38, ***** ********* ******** recognitions *** *** *** eliminate ******** ************ ***** seen ** *********.

IPVM Image

******* ** *******, ****** wearing **** **** **** recognized ** ******* ****** and ***** ********* ********** as * ********.

IPVM Image

*******, **** ******* **** a *** *** **********, subjects **** *** **********.

IPVM Image

**** ******** ** ***** issues:

** ** *** **** with *** ** *********, the ******* *** ********** of *** **** **** is ******* / ********, the **** ****** ** is **** ******** *********** biometric ****** *** ****** / ******* **** **** and ***** *********** *** matching *****. **** ******** results ** ***** ***** scores. *** ********** *** Proximity *********.

**** **** ** ********* job ** ******** **** sufficient ********* ******* *** available *** ********. **** both ******* *** ***** are ****, >**% ** high-value ********* **** ** unavailable.

Masks ****** ********

*********** ******** *** **** reduced **** ******** **** masks ** *** *****, only ******** ************ ** shallow ****** *** ***** sometimes ********* * ****** person ** * ********.

IPVM Image

**** ******** ** **** a ***** ******* **** be ********* **** ***** will ******* ****** ******* performance. **** **** **** this **** ** ** available.

Age/Gender/Expression ***********

****'* *** ********* ********* estimated ********' **** ** ~5-10 ***** ****** **** actual, **** *** ************ at *** *******'* *** or *****. **** *** not ****** *** **** of ****** ** ***** in *** "*****" ***** (<18 ***** ***) ** vice *****, ****** **** analytics ** **** ****** (e.g.,******** ********** ******).

IPVM Image

****** ********* **** ******** weighted ** **** ******, with ******-****** ******** ******** as ****** ************. **** is ******* ******* ****** ********* ** have ******.

IPVM Image

****'* ********* ********* ** not ******* * **** range ** ********, *** instead, ******* "*******" ** a *-***% ********** *****. In *** *******, **** subjects *** * ******* expression, *** ******* *** accurately ********.

IPVM Image

Aggressive **** ********* ******* **** ** ****/***** *******

**** ********** ******** ********* objects *** *** ***** of ********' ***** ** faces ** *** *******.

*** *******, *** ***** taillights ***** **** ******** as * **** ******* times:

IPVM Image

** **** *** ****** and ******* ********* ***** here:

IPVM Image

**** **** ***** ******* were ****** ***** ** "Unrecognizable" *** ***** ********** as * ******, **** a ********.

*******, *** ***** ** subjects' ***** **** ********* shown ** "********" ****** our *****:

IPVM Image

****** ** **** ***** they **** ****** ***** as **************, ** ****.

IPVM Image

***** ***** ****, **** said **** ***** ******* are ************ ***** ** give ***** ****** ******* into ********** ** *** system, ***** *** ** useful **** ** ************* standpoint (******* ***********'* ********) *** **** "**************" objects *** ** ****** hidden **** *** ***** list.

**** ******** ***** **** a ******* ******* ** view *** **** ******* through *** ****** **** some ***** “**** ***********” platforms. **** ********* ****** a ***** ** ******* that *** ** *** not ** * ****, process **** *** *********** / ********* ********, *** when * ******** ***** is *** ********** ***** their ***** ********* **** simply ******* ** *** image ******* **** ********** the **** *** ******** been ********* ** ** it **** *********** * face.

****, **** *** ***** Archive, ****** *** * view **** **** ******* with ******* ************ **** other ********* *******. ** “unrecognizable” ****** *** * concern, ** *** ** what ***** ************ ** and ****** **** *** fact **** **** *****. In ***********, ****** “******”, and **** ******** “******* Unrecognizable ****** **** ******”. These ****** **** ** long ** ******** ** the ***** ******* *** the ******** ********* ****.

~150-160 *** ****** *** ***********

*** ****** *********** ~***-*** PPF *** ******, ************* higher **** ***** ********* such ** ********* (~**-** PPF) ** ********** ****-**** camera (~**-*** ***).

IPVM Image

Versions ****

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

  • **** ******:*.*-**
  • **** ***** ** ***:**.*.*

Comments (13)

** ******* **** ******:

IPVM Image

Agree
Disagree
Informative
Unhelpful
Funny: 7

* ***'* **********. ** the ***** ** **** article *** ***** **** Facial *********** ****** ****, but *** *********** ** made ** ****** ** on ****. ** ****** only ********* ** *********. Is ** ***** **** I've **********?

Agree
Disagree
Informative
Unhelpful
Funny

*** ****** **** ** the **** **** *** is "**** ******", **** *** *****. The ****** ******** ********* and ******* **********, **** the ****** ******** *** matching. **** ** *** graphic **** **** *** marketing:

IPVM Image

*** *** *** *******, it ** *** ****-*********.

Agree
Disagree
Informative: 4
Unhelpful
Funny

***** ***,

***** - ****** *** adding * ***** ** clarification ***** **** ******. I'll *** ** **** a *** **** ** some ********* ******** *** covered ** *** **** article **** **** ******* context *** **** ******.

*********** ************ #* – with **** ****** ******* on *** ***** ********** of ** ******* ******* from ****, **** *** video ********** *** **** detection ** ******* ** the ******. ** **** how **** ********** ******** though *** ************ *** already ************ *** ****** SAFR ******** ******* ******** on * ******* ******. Cameras **** ***** ************ are ******** ** *** the ****** **** ****.

**** ** ***** ****? Reach *** ** ** directly.

** **** ********** ***** becomes ********* ****** ********** devices, **** ********* *** able ** **** ******* resources *** ** ********* data ******* *** **** the ******* ******* ****** upon **** ***, ******* that ** * *********, our *******, ********* ** even ** ** ************ camera. ** ********** ******* with **** **** ******* chipsets ****** **** ** reduce ***** ********** **** data *******, ******* ****** to **** *** ******* new *************.

**** ****** **.* ******** the ***** *** ***** in *** **** *********** process (*** **** ******** intensive *****) ******** ** cameras. ***********, ****************, *** Alerting ***** **** ** supported ** *** ****** as ******* **** ********** compute ************ *** ********. I ****** **** ****** manufacturers **** **** ******* devices ******* *** ****** later **** **** ***** on **** ** *** coming *** ** ********* manufacturers *****.

IPVM Image

*** **** ***** ** III *** *****-*** ******* chipsets ************ ********* *** AI ************, ******** ******* with ******** ************ ** provide ************ **** *** inside *** ****** *** reducing *** **** *** expensive ****** *******. ******, by ********* * ****** that ******** **** ******, our ********* **** ******* $500 *** $*** ** upfront ***** **** ********* would **** ** ***** on ******* ** *** video ********** *** **** detection ***** **** ***** performed ** * ****** in * **********. *** savings ** **** ******* when *** ******** *** monthly ***** ** **** center **** *****, ****** maintenance *** ******* *********, etc.

** **** ****** * product *** *** ***? No. **** ****** ** component ** *** ******* SAFR ********. ** ******** SAFR ********’* **** *** opportunity ** *********** ********** video ******** ********** ****** the ******** **** **** meets ***** ******** ******** needs. ** ****, ***** is ** ********** ********* costs ********** **** ***** SAFR ****** – **’* all **** ** *** base **** *******. ****, the ** ** *** camera ** **** ******** to ********** *****, ****** and *************** ***** - confirmation **** *** ****** and *** ******* **** servers (** **** ** cloud) *** ************* **** and *****. ******** ****** being **** ** ****** their ******* *** * central ******** ****** **** connecting ** **** ****** independently - ***** ** significant ** ** *** camera ** "** ******".

********* *** ****** **** on **** ******* ** Windows, *****, ***** *** now **** **** ********* for ********** *** ***** two ***** ** *** face *********** ******** – detecting ***** ** *** video ****** *** ********** the ********* **** ***** for ********** ********** (************, mask *****, ****** ********, face ***********, ***.). ********** tasks *** ********* ** the **** ******* ******, either ** ******* ** in ****’* ****** ***** service.

******** * ***** ******** with **** ***** *******… with **** ****** ******* on * ********* ******, a ******** **** ************** dispersed ********** *** ****** face *********** ****** *** of ***** ********* ******* having ** *** *** new ********. **** **** Inside ******* ** ** Axis ****** **** *** have * ********* ******* and ********* ******* ********** multiple ************** ********* ********** - ******* ****** *** additional ** **** ********. Faces *** ******** ** each ********, *** **** centrally *********. ********* ****** all ***** ******** ********** and ************** ** *** central **** ******, ***** orchestrates *** ***** ********** activities, ******* **** ** on * ********* ******, or **** ******.

*** * ***** ******, sports *******, ** ************* complex, ******** *** ******** engagement ************* **** ****** the ****-******* ****** **** at ***** *********. ********* additional, *****-******** ******* **** SAFR ****** *** ** additional ****** ************ ** their **** ****** – they *** ****** *** temporary ******* ******** ****** and **** *** ****** operation ** ******* **** a ***** ********** ** the **** ********.

**** ********** ********? **** them **** ** *** IPVM ******** *******, ** send **** ****** ** me ** *****@************.***.

Agree
Disagree
Informative: 4
Unhelpful
Funny

**** ***** ** ***** at ***, *** $**** for ** ** ******* Axis *******-**** ***** $**** on * **** ********* licence *** **** **** cloud ************ ** * split ** *** *******.🤦

Agree: 2
Disagree
Informative
Unhelpful
Funny

***** *********** #*,

***** ******** - * wouldn't *** *** **** either. ******* ***** ** some ********* ** ******* I *** ***** **.

**** ******** *** **-**** deployments *** ** ********* as ********* ** ** an ****** ************.

** * ******** ***** their ******** ** ** hosted ** *** *****, then **** **** ** annual ************ ****** ********* to ****.

** **** ******** ****** be ****** *** **** an ****** *** ***** Hosted *******.

****'* ******* ** ****, is **** **** ****** works **** **** *** SAFR ***** ****** ******** and *** **** **-**** deployment ****** - *** it's * ****** ** configure **** ******. ****** install **** ****** ***** the **** ******'* *** interface, ** *** *** Axis ****** ******* ** install **** ****** ** multiple ******* **************.

**** *********, *** **** your ******, *** **** the **** ****** *********** under *** **** *** in *** **** *** interface. ****** **** ********. SAFR ***** ****** ******** only **** ** ***** their ***** ******* ** and ********. **** ****** automatically ******** ** *** cloud *******, ********* *** camera, *** ******* ******* processing ************** ** *** camera.

**** **-**** ********* **** select ****** **** *** SAFR ******** ***. **** you ****** *** ** address *** *** **** servers *** * ***** SAFR **** ******* *** password.

*** **** ******, **** on-prem *** ***** ******, manage ******** ** ****** of *** ******* - not ******* **** ** complex ********** *** **** to ******** ******* ****** for ******** ** **** Inside **** *** **** Server.

******** * ***** *** our ******* ** **** talk ***** ***** *******, but ***** *** ******* up *******, **** * see ** **** *** the *** ********* ********* cameras *** ** *******: Axis ***** ** *** (box ******) $***, *** P3255 *** (**** ******) $739. ** *** ******* an ******* ********* *** the ***** **** ***** is ** ********** ****** there.

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

Agree
Disagree
Informative: 1
Unhelpful
Funny

** * ********** ***********? This ** **** * seem ** **** **** up **** ****** ******* with ** **** ****. Stranger ********* **** ** my ******** ****** ** it. ******* ** *** a ******, * ** considering **** *** ****. However, ***** *** * am ******** **** ** analytics.

Agree: 1
Disagree
Informative
Unhelpful
Funny

******,

***** ** **** **** you **** **** *********** SAFR. **** ******** ******** Access ******* ** * number ** *******. ******* this ** ********* ******** as * ******-****** ********, it ** **** **** expensive ** ******* **** issuing ***********, *** *****, or ***** ***** ** authentication ** *** ******.

**** ** ****** ******** for *********** ***** ****** a ***** ******** ** use *****. **'* ******** to ******* ****** ******* (physical *** *******), ********* / ************ **********, *** as ** ******** ********.

*'* **** **** ** hear *** *** ******** with *** **** ******* of "*********". **** *** briefly ******* ** ** the **** ****** ** I'll ****** * ****** further ****.

****'* **** ********* ** a *****-**** *******. *** initial **** ********* ****** in *** ***** *********** Gateway (*****) *********. *** process ******* ***** ** the ***** ****** *** sends *** ********* **** to *** **** ******. VIRGO ********* ** ***** the **** *** ****** time ** ** ******* and ** ******* ***** of *** **** **** be **** ** *** SAFR ****** **** * higher ******* **** ***** becomes *********. ** ******, this ** *** ************ by *** *************. *** VIRGO ********* *** ** deployed ** *******, ***** and **** *****, *** you ******* **... **** ACAP ******* *******!

**** ****** **.* ** actually *** ******* ******* of ***** ***** *** same **** ********* *** configuration ********** **** ***** be **** ** *** instance ** *****. *** of **** ** ******* centrally ****** ****** ************** of ***** ******* ****** - ********** ******* ** one *****.

** *** ******* **** Server, * ****** ******* assessment ** *** **** is *********. *** ********* answered ** **** **** include: ** ** ***** a ****, ** ***** sufficient ******* ********* **** available ** *** **** to ******* **** ******* / **** ******** **** recognition, ** *** **** occluded *** *** ** so ** *** ********* recognized ** * ***** style **** ** ********* other **** ** ********* true ********* ****?

*** ** **** ******** enables *** **** ******** to ******** *** ***********. For *******, ** *** face ***** ** ******** by * ***** ***** mask, *** ********* *********** workflow ******* ** *** periocular ****** ** *** face, ************ *** ****** covered ** *** **** mask.

** *** **** ***** has ************ / *** enough **** ******* ********* data ** ******* ******** face ***********, **** *** image ** ********** ** "unrecognizable". ** ******* ** this ** **** * face ** ******* ** a **** **** *** sunglasses **** ***** *** view ** ********* ****. Configurations *** ********* ** SAFR ** ****** *** quality ******* *** ************ of **** ****. ** default ** **** **** information ********* ** ********* can ** ***** ** everything **** ** ********* out ** *** ***. We ***'* ****** ** out ** *** *** deployment ****** ** **** every ********'* ***** (******'* that ** ******* ******!)

**** ** **** *** step **** "***** ***" those ********* "***** ** the ******" *** ***** automobile ***** **** *** face *********** ******* **** to ****** (***** *** any ****** ** *** FR ****** ** *** publication *** ***'** *** what * ****.) *** as *** ***** *** will ******* - ******* really ** **** **** in *** **** ** their *****. [**, ** we *** ********* **** IPVM *** ** **** mothers ** ***** **** pool ** **'** ******* more ****** **** **** one.]

**** * **** ***** is ********** ** **** sufficient ******* *********, ** progresses ** *** ****** face ******** / *********** task. ** *** ****** is **** * **** is ********** ** **** sufficient ******* *********, *** has *** ******** ** a ***** ** *** current ***** *********, **** the **** ** ********** as ********* ** * "Stranger".

**** *********, **** ** schools, **** ***** ** their ******** (** ***** of ***) ***** **** do *** ****** * stranger. ********* ** * bank ***** ** ****** cash ******** **** ***** also ** * *******. When **** ********** * stranger ** *******, ** can ******** ** ***** to *** *********** *****.

*** ********* ********* **** IPVM ********** ** * configuration ******* **** ****** SAFR ********* ** ****** the ******** ** ******* a "******** *****". ** essence, *** ****** ** the ************* ******* **** something **** **** -- "I **** * **** here **** ** *** a ***% ***** *** any ****** ******** ** the **** *********. *** you ****** ** ** alert *** **** ******* looks ************ ***** ** an ******** ****** - here ** **** ******."

[** *** ***, *** term "*********" *** * mathematical ***** ****... **'* not ******** ********* ** first ******, *** ****** that ***** **** ** learn *** ********** ** the **** ******** *** its ************* ******* ******* masters *** ******** ********* Threshold ********* *** *** it ******* ***** ** various **** *********** *********.

**** *** * ****** alerting ****** ****** ****. The ****** ***** ***** Service. **'* ** * sense ** "** **** Then ****" ******. ********* alert ******** *** ********** can ** *** ** for * ***** ***** Threshold ****** * ******** Match. **'* * ***** way ** ***** *** potential ** ***** ********* as **** ** ******** fatigue ** ******* **** to ********** **** **** need ** **** ********** and *******, ** **** they ****** **** **** to ****** *** ********* match *****.

** ** ********** ****** because *** ******** **** were ******** ** ** the ******* "************* **********" is ******* **** ******* the **** ******** ** optimize *********** ** * camera ** ****** *****. Admittedly, ** ***** ** more ** ** ****** User *******, ***** ** why ** **** **** to **** **** *** resellers. **** ********* **'* have ******** ******** ** the ******* ************** *** should ** **** ** perform *** ****** *** optimization ***** ** ****** of *** *** ****. The ****'* ******* **** is **** ********* ** assist **'* *** *** users ***** - **'** motivated ** *** *** customer's *******.

*'** **** ******** ** the *********** *** ********** of **** *********** ********** for **** ** *****. I ********** **** ** has *** ***** ** delight (********** ********* ** theme *****, *** ************ access ** ****** ********, etc.), ** **** ** the ********* ** ***** harm **** ** ** not ******** *** ** and *****. ** **** these ****** *********.

******* *** ** ********* we *** ********** ***** at ****. ****** *** time ** ****** ** systems *** ******** ********** for ******* *********** ** the ******** *********** ** which **** **** *******, as **** ** ************* where **** ******* ******* question *** ***** ***** performance ******** ***** ******, is **** ** ***** an ******* ****.

**** *** **** * good *** ** *********** some ** *** ***** where *** "*** ** the ***" ************* *** limitations - **** ** good *** ********* *********** for ***** ** *** face *********** ****** ** understand.

** *** *** ******** looking ** ******** ** FR ********, *** ***** like ** ********** ******* how ******* *** **** platform ** ** **** specific ***********, ** ***** be ***** ** ******.

Agree
Disagree
Informative: 3
Unhelpful
Funny

***'** **** **** ** in *** ****. ** will ********** ** *******. The ******** ********* ** our ****** * **** selected ***** **** ******** opportunities ** ******** *** person ** **** *** contained ****** *** *********. Playing **** **** ******** for ****** *******, * get ** **% ** 99% ******* **** ***** I ******** ** ** very *********. *** **** like ****, * ***** get * ***** ********, which ** ** ****** is *** **** *****. The **** ***** ********* is * **** **** using **** *****, *****-***, and *** ****** ***-*****. However, * *** ** targeting *********** *.* ******* this ****** *** *** how **** ****** **** play.

*** ********** ****** ** physical ******, ** **** as *** ********* ** security (***** ** *** ground), ******* ******** * great *** ** ***** your ******* ** *** campus.

Agree
Disagree
Informative
Unhelpful
Funny

** ***** **** ***** the "*********" *****.

Agree
Disagree
Informative
Unhelpful
Funny

***** *********** #*

*'* *** ******** **** the "*********" ***** *** are ********* **, *** I **** **** ***-******* builds ** **** ****** had **** ** **** - **** *** *********** experience **** * *** of *****. :-)

*** **** ******* *** SAFR ******** **** ********* a ***** ** *********** in ***** ********** ******** - ** ***** *** result ** ********* ********** great *** ********* ******* to ****** *** ******* of **-**** *******, ** enjoy *** ********** ** a ***** ****** ********.

*** ***, ** *** see ** *****, ****** report **. ** **** to ****** **** *** improve *** **** ********** for *** *** *********.

**** *** **** - it's * ***** *********** to ******** **** *** have *********** ******** ***** in *** **** ** surveillance, ******** *** ****** control.

******,****

Agree
Disagree
Informative
Unhelpful
Funny

*** **** ** ** ‘No ******** *********’.

Agree
Disagree
Informative
Unhelpful
Funny

***** ***,

**** *** ***'* ****** of **** ****** ** correct **** ******* ** SAFR ******. ******** ******, SAFR ****** ** **** one ********* ** *** much ****** **** ********, and ** *** ***** different ******* **** *** SAFR ******** ******** *** liveness *********, *** ** them *** ********* *** processing ***** ****** **** any ** ***** ******.

*** ***** ******** ****** is ******* ** *** the ******** **** ** test ******** ******* ******** needs *** ****-***** *** cases. ***** ******** **** I've ****** **** *** said **** ***** ******* stop * ****** ******* up * ***** ** a ******'* **** ** they ****** ** ************ cameras. ******, ** * former ****** ******* ******, that ****** ***** ** sufficient ****** ** ******* contact ** * ******** security ***** ** *** enforcement *******, ********* **** the ******** ***** ** occurs.

***** ******** ** **** requested ** **** *** customers *****, *** ***** we **** ******** *** existing ************, ** ** validate ******** ****** ** access ******* ***** - i.e. **** * ****** will *** ***** **** to ****** * ******** door, ** **** ******* access ** ****** **** as ** ********** ******, such ** ** * casino **** ******* ** walking ******* * ****** transit ********* ***** * kiosk *** ************* ********* a ********** ***** *** decrement ***** ******* *** that ****.

@*** - * ***** love ** **** **** about **** ***** *** validating ******** ** * standard ************ ***** ****. Especially *** ********* ******** as ** ** **** a ******** **** *** meet **** *****.

SAFR's ***** ******* *** ******** *********

*** **** ******** **** RealNetworks ******** ***** ******* for *********** ********, *** two ** **** *** available ********* ** ***** belong ** * **** person **** *** ***** is ****** **** *** IP ************ ******.

Passive ******** & ************ ****** *********: SAFR was selected by the US DoD to advance the state of Liveness / PAD when all that is available is a single RGB stream. SAFR's approach supports video coming from USB web cams, IP surveillance cameras, and embedded camera components such as those that are typically added to ATMs / ATRs.

Depth ******* **: SAFR supports liveness utilizing data available from depth sensing camera sensors such as Intel RealSense.

Challenge / ********: SAFR supports simple challenge response mechanisms such as look left / right and smile to determine if the face in the video follows those instructions.

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

This article is part of IPVM's 7,324 reports and 971 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