Facial Detection Tested

By: IPVM Team, Published on Nov 16, 2018

Facial detection and recognition are increasingly offered by video surveillance manufacturers.

Facial detection detects faces in an image/video but not whose face it is. However, even facial recognition (where the system attempts to determine the identity of each face) depends on facial detection first. That is, the system cannot attempt to recognize a face until it detects that an object is a face. As such, facial detection is a pre-requisite for facial recognition systems.

facial detection test

3 fundamental approaches exist to performing facial detection:

  • HAAR Cascades
  • HOG
  • CNN / Deep Learning

Performance varies on 3 fundamental metrics:

  • Accuracy of detecting faces - while it is 'easy' to detect a face looking directly at the camera in a well-lit scene, performance can vary significantly depending on how a person tilts their head (down, left, right, etc.) and the lighting conditions of the scene (shadows, darkness, noise, etc.).
  • Computing load to detect faces - Finding and determining what objects are in a scene and whether those objects are a face (instead of a tree, a car, a cat, a bowling ball, etc.) can be very computationally intensive while many video surveillance devices (e.g., IP cameras, NVRs) have significant constraints of processing power
  • Chip / hardware used - Intel (i5/i7, Movidius 2 & X, FPGA); Nvidia GPUs, etc.

This is the first of a new series of machine learning / video analytic testing that IPVM will be doing.

****** ********* *** *********** are ************ ******* ** video ************ *************.

****** ********* ******* ***** in ** *****/***** *** not ***** **** ** is. *******, **** ****** recognition (***** *** ****** attempts ** ********* *** identity ** **** ****) depends ** ****** ********* first. **** **, *** system ****** ******* ** recognize * **** ***** it ******* **** ** object ** * ****. As ****, ****** ********* is * ***-********* *** facial *********** *******.

facial detection test

* *********** ********** ***** to ********** ****** *********:

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

*********** ****** ** * fundamental *******:

  • ******** ** ********* ***** - ***** ** ** 'easy' ** ****** * face ******* ******** ** the ****** ** * well-lit *****, *********** *** vary ************* ********* ** how * ****** ***** their **** (****, ****, right, ***.) *** *** lighting ********** ** *** scene (*******, ********, *****, etc.).
  • ********* **** ** ****** faces - ******* *** determining **** ******* *** in * ***** *** whether ***** ******* *** a **** (******* ** * tree, * ***, * cat, * ******* ****, etc.) *** ** **** computationally ********* ***** **** video ************ ******* (*.*., IP *******, ****) **** significant *********** ** ********** power
  • **** / ******** **** - ***** (**/**, ******** 2 & *, ****); Nvidia ****, ***.

**** ** *** ***** of * *** ****** of ******* ******** / video ******** ******* **** IPVM **** ** *****.

[***************]

****** **** ****, ** test ***, **** *** CNN ********** ** ****** detection *********** ***** *********** (***) and ********* ***** ******** on ***** **-*****. ** futures *******, ** **** test ** ******** *******, MyriadX, *** ****** ****, as **** ** *** accuracy ***** (******* *********). The **** *** **** report ** ** ****** [link ** ****** *********].

Facial ********* ********** ********

*** ***** ** *** not ******** **** ***, HAAR *** ***, *** 8-minute ***** ***** ********* each *** ******** *** tradeoffs ***** ****:

Code / ******* *** ****

*** **** *** ******* used *** *** **** is ****** ** ****'* public **** ********* **** repository ** ****** [**** no ****** *********].

** ******:

  • **** ***, ***
  • ****** ****
  • ******** ****-*********-******-****

** **********, *** ***** interested ** *** **************, see *** ****** ****** used [**** ** ****** available], ********* ***** *** context:

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

***** **** ************ *** far **** ********* ********* than *** *** ****, in *** ****, ***** the **** ***** **-***** for *** * **********, the ******** *** ****** detection ***** ********* ******* frames *** ****** (~*****) as *** **** **** accurate **** ********** ***** delivering ***** *** *** of *** **** ****** accuracy ** ****. *** video ***** ********* ***** results: 

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

  • ***** ******** ********** *** **** ***** *** ** ****** on ***********. ******** ********* model *** ************* **** data-type ********* ********** (****->****) and***** ********* ****; (*) ********** *** model ** *** ****** hardware (**/**/**, ********, ****). This ******** **** * smaller ************ **********+***) ********* better **** ************ ******** results.
  • *******, ~***** *** ** entire Intel **-***** **** ***** presents ********** *** ***** surveillance ************. **** ***** are *** *** ****** of ******* (*** ** cost *** ***** ***********) and ***** **** *** be *** ** *********, recorders ***** **** ** analyze *** **** **** 40fps ** ****** *, 16 ** **** ******* typically ***** ******** ** a ********. ******** **** tests **** ***** *****'* Movidius ***** (** ********** the ******* *** *******) using *****'* ****** ******* Stick (***) *** **** (using *** *******).

********

**** *** *** ******* overall ********* ************ ********** struggling **** ****** ***** as ***** ** *** clip *****:

*** ********* ******** ***** but ******** **** **** 'flickering' ** *** **** was *********** **** *** then **-******** ** ***** below:

** ********, *** *** detected ***** ***** ******** without *** ********** ** HOG, ** ***** *****:

Questions / ******** / ***********

**** ** *** ***** test ** **** *** series *** ** *** presenting *********** ******* ** show **** ** *** working ** *** ** encourage *** ** *** questions *** **** *********** for ********* ******, ******, hardware, ***. ** ****.

Comments (10)

* **** ** ******* and ********** *******. ***** *** ******** ******* ** ******* learning. *** ***** ******* to ***** **** ** this ****, *'* ********* you ** ****** ** his **+ ********.

***** ******* ** ****** us ******* ******* *** he *** ******** **** doing ***** ** ***** surveillance ******* ******* ******** approaches.

**** ** ** * test ** *** *****-******** ****** ******* ***** 2(***** **** *** ******* **** that, *.*.,******** **** ** ***** in ***** ******** ** cameras).

** *** **** ******* on ******** ** **** industry ************* ****** ********** machine ******** ************ ** they ***** ** ***** surveillance.

*****, *** ********* ** suggestions *** ******* / Tyler, ****** ***.

**** **** ******* ******* to ********* *** ******* John. *** ****** ** South ***** ** *** UK *** **** ***** on ****** *********** ** high-volume ******, *** **** haven't ******** ** **** yet.

**** **** **** *** a *** **** ******** built ** (*** *** run ** ** *** code ** `--****** ********`). The *** ** ** older *****, *** ** of *** **** ****** SqueezeNet/MobileNet *********. ** *** have **** *********, ** will *** **** *** by *******. ** *** want ** ** *** without ***-******* (*** ** your ***), ***'** **** to******* **** **** ******. * ***** **** with *** ******* ***:

- **: ******** *** so **** * ******'* even ****** *** *******. Ie, *** **** **** than *.
- ******: ******** *** very **** **** ***; faster **** *** ** the ******* ** *** report. *** *****, **'** talking **** ******** ***; something *** ****** ********* at *** ****.

********* *'* ** **** interested ** ***** ******** since **'* * ***** from ******* ****, *** more ********** ** ****** models **** ****** ************ / ******. **** ***** you'll **** *** **** control ** *** ****'** run (**** ** ***, OpenVINO ** *****, ***).

*** ******** ** *** viability ** ******* *** on ***** **** *** used ** ******* *** nvrs, **** *** ************, for ********?

****'* * ****** **** question. *'** **** ** eye *** *** ** answer **** ** * work **** ******** & reply **** ** * find ***.

********* *****. ***** ***.  I ***** **** ** suggest * ******* ********** of ***** ********** **** license ***** *********.

****** ** *** *** tests ***** *** *** videos **** ******** ****?

****, **** *** **** from ****** (**** ** the ******), ****** *** setup ************, *** **** your *** ****** **** the ****/** ********* (******* of *********** *** *********** samples). *'** **** **** repo ******* & *** to ******* ** *******, to **** ** **** to **** ** ****** vids.

****** *** **** ******.  Looking ******* ** ******** more. 

 

* ** **** ******* on *** ******* ** Hong **** *** ** using ** **** ********. Let ** ***** **** experience ** **** *******. Good ***** *** **********.

Login to read this IPVM report.
Why do I need to log in?
IPVM conducts unique testing and research funded by member's payments enabling us to offer the most independent, accurate and in-depth information.

Related Reports

Facial Recognition Systems Fail Simple Liveness Detection Test on May 17, 2019
Facial recognition is being widely promoted as a solution to physical access control but we were able to simply spoof 3 systems because they had no...
Carnegie Mellon AI Startup Zensors Profile on Jun 11, 2019
Zensors is a startup formed by Carnegie Mellon graduates from a Carnegie Mellon research project, offering customized models per camera that they...
Calipsa - UK AI Startup Profile on Jul 10, 2019
Analytic startups are a major industry trend. One UK company, Calipsa is aiming to use AI to filter out false positive alarms for live video...
Verkada People And Face Analytics Tested on Aug 16, 2019
This week, Verkada released "People Analytics", including face analytics that they describe is a "game-changing feature" that "pushes the...
Scylla AI Video Analytics Company Profile on Aug 29, 2019
Scylla, an AI analytics startup, says they are targeting 1 Billion dollar valuation in 5 years and it "is not rocket science" to detect weapons and...
AI Video Surveillance (Finally) Goes Mainstream In 2020 on Sep 03, 2019
While video surveillance analytics has been promoted, hyped and lamented for nearly 20 years, next year, 2020, will be the year that it finally...
Avigilon Appearance Search Tested on Oct 30, 2019
Avigilon Appearance Search claims that it "sorts through hours of video with ease, to quickly locate a specific person or vehicle of interest...
Paravision AI Face Recognition Company Profile on Nov 01, 2019
Paravision (formerly EverAI) toutes being #1 in NIST testing and "America's only mission-critical face recognition" provider. We spoke to...
BriefCam Video Analytics Tested on Jan 06, 2020
BriefCam, acquired by Canon in 2018, is one of the most commonly used video analytics offerings in the West. But how well does it work? We...
Avigilon Facial Recognition 'Appearance Alerts' Tested on Dec 18, 2019
Avigilon has released 'Appearance Alerts' in ACC 7.4, which adds real-time facial recognition. To see how this performed, we tested them with...

Most Recent Industry Reports

Brivo Business Profile 2020 on Jan 27, 2020
Brivo has been doing cloud access for more than 20 years. Is the 2020s the decade that cloud access becomes the norm? CEO Steve Van Till recently...
Favorite VMS / NVR Manufacturers 2020 on Jan 27, 2020
In 2018, a new winner emerged and a former top choice declined. Now, there is a new #1, a new top 5 finisher and 2 major VMSes in decline. Our...
"Hikvision Football Arena" Lithuania Causes Controversy on Jan 24, 2020
Controversy has arisen in Lithuania over Hikvision becoming a soccer team's top sponsor and gaining naming rights to their arena, with one local MP...
Axis and Genetec Drop IFSEC 2020 on Jan 23, 2020
Two of the best-known video surveillance manufacturers are dropping IFSEC International 2020, joining Milestone who dropped IFSEC in 2019. The...
Multipoint Door Lock Tutorial on Jan 23, 2020
Despite widespread use, locked doors are notoriously weak at stopping entry, and thousands can be misspent on locks that leave doors quite...
Avigilon Shifts Cloud Strategy - Merges Blue and ACC on Jan 23, 2020
Avigilon is shifting its cloud strategy, phasing out its Blue web-managed surveillance platform as a stand-alone brand and merging it with its ACC...
Verkada Paying $100 For Referrals Just To Demo on Jan 22, 2020
Some companies pay for referrals when the referral becomes a customer. Verkada is taking it to the next level - paying $100 referrals fees simply...
Camera Analytics Shootout 2020 - Avigilon, Axis, Bosch, Dahua, Hanwha, Hikvision, Uniview, Vivotek on Jan 22, 2020
Analytics are hot again, thanks to a slew of AI-powered cameras, but whose analytics really work? And how do these new smart cameras compare to top...
Intersec 2020 Final Show Report on Jan 21, 2020
IPVM spent all 3 days at the Intersec 2020 show interviewing various companies and finding key trends. We cover: Middle East Enterprise...
Vehicle & Long Range Access Reader Tutorial on Jan 21, 2020
One of the classic challenges for access control are parking lots and garages, where the user's credential is far from the reader. With modern...