PureTech Video Analytics Examined

By Brian Karas, Published Apr 21, 2017, 01:11pm EDT

PureTech's analytics were chosen for a US border protection system (see related post), which the company claims no other analytics vendor was able to meet the requirements for. PureTech is using deep learning techniques in portions of their products, and claiming the ability to detect targets at ranges up to 7 kilometers with higher-end FLIR cameras.

In this report we examine PureTech's technology, market strategy, and pricing in comparison to other analytics providers.

PureTech ******* ********

******** *** ******* ** 2005 *** ** * privately ****/********** ***** *******. The ******* ******* ** software *********** *** ***** analytics ******** ** **** on **** ********.

******* *** ***** ****** are ******** ** ********, primarily ** ***** *******, AZ ************. ***** ********* held, **** ***** *** disclose *******, *** ** would ******** ** ** be ******* $** ********.

******** ********

*** ********* ********* *** on * ****** ********, using **** ********. ******** does *** ***** **** in *** **** ** appliances ** ***-****** *******, preferring ** *** *********** supply *** ****** ******** for *** *********, *** instead **** *********** ** selling ********.

* ****** ** ****** analytics *********, **** ** object **************, ******* ******* detection, *** ***** ********* are ********* **********'* ********* ********. *** ******* ****** to **** ************ **** most ****** *** *********, but ********** ***** *** *** (********* ***** *****), ** the *** ***** ************ often **** ******* *** advanced ******** *** *************, such ** **** **** of ****** *********.

** ******** ** ********'* analytics ***** *** ******** is ***** ** *** video *****:


Geospatial *********

******** ******** ********** ******* of *******, *** **** this ******* ** * prime ************** *** ***** software.  **** ***** ** offer * ***** ** precision *** ********* ******** that other ********* ******* *** unable ** ****, ** they ******* ********** *********** at ***. **** *** be **** ** **** targets ** * **** map, ** ******* * PTZ *** ****-******** ** detail ******* *** *******.

Deep ********

** ***** ** ********** detect *******, *** ****** non-threat *******, ******** ********* and ******* ***** *** deep ******** ****** *******. This **** ******** ************ is **** ** *** object ********** ********, ** well ** *** ********** positioning ******, ***** ** used ** **** ** their ***********.

*******

********* ** ********, *** majority ** ***** ********* functions *** ******** ************, with ********* ********* ********** functions/features ******** ** ***** requirements. ***-******* ******** ***** around $***, *** *** go **** ****** ** customers *** *** ******** analytics ********* ** **** channel. ********* ** ********, this ******** ** ********* allows **** *********** ** system ******. ***** ******* costs *** *** *** software ****, *** ** not ******* ****** *** cameras, ** ****** ******** required ** *** *** software.

Target *******

******** ****** ******* ***** end *****, ******** ************** facilities, ********, *** ******* organizations **** **** ** have **** ****** *****, and *******, **** ** comes ** ********* **********. Often ***** ***** ********* have **** ********** *****, spanning ********* ** ****.

Typical ***********

** **** ***********, ******* cameras *** **** *** their ******* ** **** without ********, *** ** see ******* ******** ******* conditions. ****** *********/************** ********* are **** ** *** video ***** **** *** thermal *******, *** **** PTZs *** ************ ******* to *** **** ***** a ****** ****** ** detected ** ******* ********** visual ****.

*** ****** ******** ********** the ********* *** ** centrally *******, ** *********** in ** ****-**** ************, using ***** ****-****** ***, such ** ** ***** ***.

Border ******* 

*** *** **** **** on *** ** ****** project *** ***** *** of ********.

***********

***** ***** *** ******* companies ******* ********* ******** to ********* ********** (********,*******,**** (***),********), *** ***** ** offer *** **** ** target *********** ****, ** ranges **** ***** ********'* claims. ***** ******* **** tend ** ** ***** cost ***-*******, ********'* ********* are *********** **** **** their ******* (****** ******** cameras **** ** ** more expensive **** ************), *** license ***** *** *** other ******* *** ********* less ********* **** ********'* offerings.

*** ********* **** ** mention ***/***********/********** **** *** ******* ****** *** **********, ******* ** **** not ****** ****** ******** of ***** ********, ** focused ******** ******* ** location ******** *** ********* can *** ******* *** they ******* ** ******** in **** ******.

 

Comments (11)

It's a joke? I'm sure is running windows 95...

Jordi, do you mean the user interface shown in the video? If so, I think it looks dated as well. Please clarify.

Btw, try to make your comments clearer and more professional (i.e., calling it a joke).

Yes, I mean the user interface.

 

Btw, when I say "is it a joke?" I'm talking about my natural reaction to a dated UI, and about the PureTech video. It doesn't show anything, nor technology, nor hands on... nothing. Funny, coming froma a manufacturer claiming a technology that no other manufacture can match.I'm trying to be professional, but the video...

 

Nice weekend!

 

PS: I suspect you think I was calling a joke your post. It's not. Sorry about the confusion :) 

Jordi, I realized you were calling PureTech / the video a 'joke'. I just think there is a less caustic way to do that. Again, I agree the UI in the video is quite dated.

The GUI sample image came from PureTech's website. They sent me updated images, which I swapped into the report to show a more current interface.

What i'm interested for a change is the analytic companies will have more "challenging" demo videos that shows their real abilities to detect while ignoring false alarms.

Most of the videos are always clear days, zero wind with static backgrounds or detection area AKA no trees, animals, heavy rain/snow, flying objects/trash, windy conditions.

Many analytics that I've seen started to fail actually where the pole the camera was mounted on started to shake and vibrate on windy or heavy weather conditions... i'm curious to see performance in those situations.

Above hit the nail on the head, "never a super wind driven rain" or a 40 foot pole with a  P T Z on it, its always perfect weather, and the world is not perfect...

So "Deep learning" is the current buzz word in video analytics? How exciting!

It's not a buzzword so much as a progression of technology allowing for advancements in what can be done with video analytics.

Much like H.264 was ratified as a standard half a decade before it became widely used in security cameras. It took other refinements to H.264, and microprocessors, to make it practical to use H.264 for security video. Deep learning is following a similar trend, with advancements in frameworks (such as the recently released caffe2) and processors that are enabling it to replace older video analytics systems with less efficient or less capable methods.

Sorry for my skepticism Brian, when I hear "Deep learning" I also hear "Neural Networks" and other marketing terms thrown around liberally by some infamous analytics providers.

 

PureTech has been around forever, spun in and out of FLIR acquisition of ICX. I hope its not another Boeing debacle but from what I have seen there is nothing new under the sun.

https://youtu.be/n8IF_x5QWJc

13 years old technology 

Read this IPVM report for free.

This article is part of IPVM's 6,728 reports, 907 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports