HD Camera / Face Rec / Door Controller All-in-One

Published Apr 30, 2014 04:00 AM

Is it a security camera, face recognition reader, or a door controller?

Stand aside people: BioCam 300 claims to be all three, and more.

Should you consider this device for your systems? What are the downsides?

We take a look at the key features, pricing and competitive positioning inside.

********

********* [**** ** ****** *********] ****** to ** "*** *****'* ***** ** ** camera **** ******** ****** *********** *** display ** * ****** ****" *** if *** ****, ********* *** *** form ****** ******. *** ~*"**"**" **** essentially is * *****-****** ******-***** ******, **** large ** *** ****, *** * 7" *** ******* ******* **********: 

*** ************* ****** ****** ** ***** ************ ******, * ****-*** ****** reader, * ****** **** *******, *** a ********** **** ********** ** * single ****. **** *** ***** **** to ***** ******, *** ****** ******** surveillance ****** ** *** ***** 

*** ********

  • **** ******: *** **** ** ******** ** be ******* **** ***** ** ** can ******* ******** ** *****, *** then ******* ***** ***** ******* *** database ** '****** ********' ***********. ** to *** ***** *** ** ****** in *** **** (** ********* *** user), ***** ** ****** ** ***** and ******** ** **** **** * seconds. *** **** ****** ** *** scan ** ** ** **** ****, and *** ******** ** **** ********* lighting **** ** **** *********.
  • ** ************ ******: *** **** **** ******** * 1280 ****, ** ***, ***/***** ****** that ******** *****. *** **** *** be ********** ** ****** ** ****** locally ** *** ******, ******** ********/******, ***** *** **********, *** **** H.264.
  • **** **********: ******, *** **** *********** ******** at *** ****, *** *************** ***** *** ***** *******, *******, or *** *******. ********* ** ******** to **** ** * '**********' ****** system, *** ***** **** ********* *** credentials *** ** ****** **********, *** unit ****** ** ********** **** ****** 3rd ***** ****** *********.
  • *** ***: ***********, **** *** *** ******* included ** *** ****, ** **** not *** ***** **** ********, *** instead ******** * ***** @ *.** power ******.

****

***** *** ****** *** ** ***, it ** ******** ** **** * ~$1,300 ****** *****. *** ********************** ******** *** ****** ** *** customers, *** *** ******* *** ********** be ********* ***** ********* *****, *** different***********.

*** * **** **** **********, ***** and ****** *******, *** ***** ** relatively ********** ****** ***** *** **** of ***********, *** ** ****** ** more ** * ***** ***** **** entry ******. ** *** ******** / residential ****, ******* ** '***' - * **** *********** / doorbell *******.

***************

***** *** **** ** ******** ** indeed ***** ****, ** **** *** mean *** ****** **** ** *** any ***. ****** ** ****** ******* and ************ ** *******, ********* ****** should ** **** ** ***** ******:

  • *****:******* *** *********, *** *** ************ ZKAccess ***** ***** ********** ******* ***** [link ** ****** *********] *** ****** ******* S ** '********* **************'. ***** *** risk **** ***** ******, *** ****** approach ***** *** ********* ** * surveillance ****** **** *******.
  • **** *** ********:*****, ******* ******* *** **** *** facial********************** *** *****, ****** ****** ** face ***** ** *******. ***** ******* documentation **** *********** ****** ** *** faces ** ****, ***** ** ** indication ** *** ******** ***** ******** are **** **** * *** *****.
  • ******* ***********:***** ********* ** ***** **** * tank, **** ******** ******** *** ***** size, ** ** ******* ******* *****, water *********, *** ****** ***** *****. Apart **** ****, ************** ** ********** components ** *** ********, *** ** one **** **** ***, *** ****** device **** ** ********.

Unit *********** / *********** ********

*******, ** ****** *** ************* ** usefulness ** *** *******, *** ******* weakness ** *** **** ************** ** the ********* ** *******: ******** ********. 

*** *******'***************** *** ******** ********** *****:

*** ****** **** ** ******* *** enough **** ** ******** **** ****** of *****, *** **** *** ***** it ********** ** ********* ** *****. Other ************* **********, **** *** ******** of *** *** ** ***** ******* accessories **** ****** ** **** *** ultimately*********** **** **********. ******* ******** ******** ** *** ********* side ** *** **** ****** **** wiring ********** ** ******* *********, ******* even ********.

*** **** ** ***** ** ********'* ********* ********* *****? ****, *** ********* ******* ** ******* ****** while ******* ***** *********** **** ** user ******. ******** * ***** ** punching ** * *** *** *** be ****, *** ** ******* ********* to ***** **** **** ** ******** over *** ****** ** ***** ** you ***** ***** **.

Comments (10)
Avatar
Skip Cusack
Apr 30, 2014

If the system is running in a verification mode (1:1) as it should for access control applications, then the number of faces in the database will have no bearing on matching accuracy and false alarm rates.

RW
Rukmini Wilson
Apr 30, 2014

... the number of faces in the database will have no bearing on matching accuracy and false alarm rates.

So are you saying if it doesn't work with ~400 faces don't expect it to work any better with ~40? That's too bad because the latter is probably closer to the SOHO sweetspot...

Avatar
Skip Cusack
Apr 30, 2014

I may have expressed myself poorly. While running in verification mode the number of users enrolled in the database will have no bearing on 1:1 verification matching performance, because the user presumably asserts his/her identity and the live image(s) is compared to the database image(s) corresponding to that user only. The number of images captured during runtime and enrollment for each user will likely have a bearing on matching accuracy. However, the technical matching performance for each user may either improve or degrade with the nature and number of images used depending on how the system makes use of them.

U
Undisclosed #1
Aug 07, 2015

Skip - I'm not sure I totally follow your post. Are you saying that facial ID should be used as a sort of secondary verification mechanism, instead of being the primary/only verification mechanism?

Avatar
Skip Cusack
Aug 07, 2015

With regards to FR accuracy: The product does have provision for wiegand inputs, so it could run in a two factor verification mode: User asserts identity with a possession factor (e.g., card) and FR camera verifies identity with biometrics (1:1 match). The accuracy of this mode will be much higher than identification mode (1:Many). In ID mode, the user looks at the camera, and his/her face is compared to every other image in the database (up to 400 users @ several images / user). Achieving good technical performance in this mode requires much stronger biometric performance as the probability for false match is much higher.

U
Undisclosed #1
Aug 07, 2015

What I don't get about the recent burst of face-recognition access control products is where the real application is.

It seems like there are two primary cases:

1) Face-as-credential. This one still seems way to prone to false reads and delays, negating the ease-of-use factor when users get locked out or spend 10+ seconds at the door trying to get through.

2) Face-as-secondary credential. This one seems more practical because you're comparing the face read at the door to a known set of profile images, but it's still generally more costly and time consuming that something like an iris scan or fingerprint scan. The hardware is also bulky and ugly (for a door ornament) relative to other options.

I can't help but feel like facial recognition access control and 3D printers have a lot in common as that "gee whiz" thing that just can't really find a feasible fit in the marketplace just yet. This limits it to novelty and niche applications, which slows down overall adoption and R&D.

Are you seeing facial recognition applications that are truly viable and practical for a mainstream deployment?

Avatar
Patrick Hart
May 01, 2014
IPVMU Certified

Shouldn't the door controller be on the secure side of the door?

km
kenneth macphee
May 01, 2014

What if I take a full sized picture of an authorized person and hold it up in front of my face? At least it would be cleaner than chopping their head off!

Avatar
Lynn Harold
Aug 07, 2015

Ken - I actually tried that, didn't work. I am guessing the facial geometry stored as the credential doesn't work when you have a flat 2D image.

U
Undisclosed #2
Aug 07, 2015
IPVMU Certified

What if the spoofed face was 3d?

Shades of Brian's Frankenfinger...;)