NICE Suspect Search Examined

By: John Honovich, Published on Sep 25, 2014

Did you lose your child here?

Or here?

Now, many might say you were a terrible parent but not NICE Systems. NICE's new Suspect Search [link no longer available] aims to solve this.

In this note, we break down how it works and what the problems with it are.

*** *** **** **** child ****?

** ****?

***, **** ***** *** you **** * ******** parent *** *** **** Systems. ****'* *** ******* Search [**** ** ****** available] **** ** ***** this.

** **** ****, ** break **** *** ** works *** **** *** problems **** ** ***.

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

*****, ***** ***** *** minute ********* ***** *** the *****:

Suspect *********

***** **** ***** [**** no ****** *********] ***** searching ** ***** ** 3 *******, **** **.

**** #* ** ** create ** ******, ******* from ******, ****** *****, cargo ******, **** *****, body *****, *****, ***.

**** **** **************. *******, **** ***** marketing ***** ******* *** weak:

** *** **** *** correct *******, **** *** can ** * ********* search **** ****:

*******, *** ******* *** still ****, ** ***** by ***** *** ********* demo:

*******, *** *** ******* you ** ***, **** will ******** ** ******* on * *** ******* when *** ***** *** subject *** **** ****:

[UPDATED] ********* **********

**** ******** *** ********* tech *******:

  • **** ********* *****, **** pulls *** ******* ******. Those ****** *** **** analyzed ***** ** *****, texture *** ***** ********, enabling ****** ** ** compared *** **********.
  • *** ******* *** ***** to ** ******** ** ~10 ****** ***** ** their ******* ******** ***** CIF **********. ** ****, the ******* *** **** moderately **** **** ** best.
  • *** ~*** *******, **** would ******* ** ********** server **** ** ***** (2 ****, * ***** each) *** **** ***.
  • **** ******** ***** ****'* video ********** ******** / system *** ****** ** 'added **' ** *** party *****.

Issues ********

***** **** ********** ** really **** ******* **:

  • ******** *********- ** ** ** dark ** ****** ** there *** *******, ******** accuracy ********* *************. ** course, ***** ********** *** common ** ***** ************.
  • ****** *********** *********- ******* ****** ** how **** **** *** mounted, *** *** **** are **** ******, **** tilted, ***. **** ******* varying ***** ** ****** (from ******, **** *** front, **** *** ****, from *****, ***.). ******** accuracy ****** ********* ***** is *********** / ****. 
  • ******- *** **** ****** *** the **** ******** (*****, walls, ***.), *** **** 'suspects' *** ******** / blocked, ****** ** ****** for *** ****** ** match / ********* ****.

**** ** *** ********* are '*****', ***** *** fundamental ******** **** ****** be ********.

*******

*****'* *** **** ****** offering **, ********* *** actively ***** ** **** (e.g.,******* ******** **** **** ** 2008*** *** **** ******* claiming ****).

***** *** ******* **** large *** ***** *** would **** ** ** able ** ***** **** a *****, * *********, a **** ******,**** *****, ***.

** ****, ** ***** this ** ********* **** will ******* ******** ** NICE, *** ********* **** buyers ****** **** ********* test ** ***** ** determine ** ** *** work **** ****** ** be ****** ** ***** environments.

****** ****

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

Comments (21)

On top of everything else, the video has bad grammar! "...what if it 'was' here?" should have been "...what if it 'WERE' here?" There were other errors as well. No excuse, really.

Marketing Communications (MarComm) people tend to be idiots. They have but one job: Proofread someone else's copy, and arrange booths at trade shows. They seldom do well at either task.

Was your Typo deliberate?

It was not...but I blame Swift Key. It has been fixed.

I find your myopic view of MarComm people as ridiculous and offensive. You either have never been exposed to the true power of an effective marketing team, or you lack the business acumen to understand the value of marketing, or both.

This is punctuated by your statement that "They have but one job:" then go on to describe two different functions. Neither of which even provide an effective partial list of what MarComm performs.

I feel sorry for you.

I didn't post anonymously. The guantlet is thrown.

NICE is going to be doing a live demo of this system at their ASIS booth. That should be entertaining and an interesting challenge!

We'll be checking it out and recommend you do as well. To give it a good test, pass by their booth first, walking around for a few moments to make sure it gets you. Then come back a few hours later, have them do a search and see how well it picks you up.

We spoke with NICE today and have added some tech details, copied below:

  • When analyzing video, NICE pulls and detects people. Those people are then analyzed based on color, texture and other features, enabling people to be compared and contrasted.
  • The maximum FoV width to be analyzed is ~10 meters based on their current analysis using CIF resolution. As such, the cameras can have moderately wide FoVs at best.
  • For ~100 cameras, this would require an additional server with 12 cores (2 CPUs, 6 cores each) and 32GB RAM.
  • This requires using NICE's video management software / system and cannot be 'added on' to 3rd party VMSes.

An end user asked about tracking a suspect across different locations to see where else the suspect had been previously.

The big challenge is time frame. If you search across many cameras over a week or a month, you'll be faced with tens or hundreds of thousands of images to compare against. Lots of other people will 'look' similarly to the analytic and it will be hard to find the actual person. Worse, since the suspect may not have been to any location or only one out of a hundred of them, an operator may become very frustrated going through massive amounts of mismatches.

Oh, and you have to buy NICE recorder / analytics for every location...

"If you search across many cameras over a week or a month"

If your search spans more than 24 hours the suspect is unlikely to be wearing the same clothing arrangement, which is 90% of the basis of this kind of search.

Even someone removing or adding a hat or jacket (intentionally or just from going indoors/outdoors) will easily thwart this type of appearance search.

It sounds nice, and in controlled environments it demos well, but it will have very limited practical use.

I thought 3vr claimed they could do this sort of bigbrotheresque stuff too.

I found the Schlage wireless lock that offered WEP support to be appalling too.

3VR is probably the closest comparison. I don't think the two are claiming the exact same approach, but there are similarities in the overall search / find a needle in a haystack approach.

In some of these scenarios, I think you'd be better off using facial surveillance / search (like 3VR) because, while matching faces from uncontrolled video is hardly perfect, it is a lot more accurate that matching bodies.

Btw, I reached out to 3VR on Friday for their take on NICE's approach and their best comparable offering. I'll update if / when I hear from them.

Even if it worked in a mall say, are malls asking for this service? I don't see any good ROI for this - try charging the distressed parent for the service - see how that goes over...

Well, I am sure lost children and crime suspects are an issue for malls.

However, to your point, what is the return / value of being able to find them faster?

I suspect NICE will target this more for critical infrastructure, which is their core market anyway - airports, city surveillance, etc.

"However, to your point, what is the return / value of being able to find them faster?"

This would be a time-saver for the mall, and in essence a deterrent to abductions or certain crimes. Many VMS vendors already tout various features that expedite searching video, this is no different. I don't think anyone would propose a mall charge for this service, but if a parent claims a lost child they are going to be compelled to put high effort on finding the child. You could easily suck up 20-40 man hours in a short time hunting down a child reported missing. Of course, my personal opinion is that this avatar-building video game will just suck of time in a different way and produce bad results, the point is that the value is faster response and resolution.

"Many VMS vendors already tout various features that expedite searching video, this is no different."

Specifics please, since you are criticizing a competitor and since your counterargument is the vague 'expedite searching video'

Every VMS demo I've ever seen has tried to highlight something about their search capbilities.

Avigilon lists it promienetly on the ACC product page:

http://avigilon.com/products/video-surveillance/avigilon-control-center/

Briefcam has built a whole company around video search.

Exacq issued a press release about their video search enchancements: http://d1ni7hpbick8ut.cloudfront.net/exacqVision-6.0%20Release_final.pdf

Genetec/AgentVI tout search enhancements: http://www.genetec.com/solutions/resources/agent-vi-video-analytics-solutions-integration-with-security-center-and-omnicast

OnSSI demos search heavily, and has a quote calling out the search function: http://www.onssi.com/products/ocularis-overview

Not sure exactly why you took exception with me saying that NICE is basically trying to do what every VMS does: highlight something special about their video search functionality. I'm not criticising it anymore than you have already, I'm agreeing with you in saying that I don't think it will live up to the hype.

Come on, you can't seriously be comparing Exacq and OnSSI search 'enhancements' to what NICE is claiming with suspect search. For example, here's Exacq's big search breakthough catch-up.

NICE is saying they catalog individual people and can rank those people by similarity.

Is anyone else on your list claiming anything like that?

I obviously am skeptical about NICE's ability to deliver this, but you can't fairly cut it down by comparing it to vanilla VMS search.

The OP to this thread said "I don't see any good ROI for this - try charging the distressed parent for the service".

You replied "However, to your point, what is the return / value of being able to find them faster."

What I was saying is that this 'feature' isn't something that NICE is proposing (from my understanding) that you'd charge money for, etc. It's a search enhancement feature to try and make it easier/faster for operators to search through video.

Every major VMS vendor seems to call attention to particular aspects of their search capability. I also think it is fairly well known that operators normally spend at least as much time looking at recorded video as they do live video, and often times the trigger to look at recorded video is some event where they also want to backtrack or collect related frames of video.

I'm NOT saying that NICE's avatar search is in the same category as a basic pixel-change search or motion-event search.

The ROI to this is operator efficiency, not new revenue streams.

I can't see the business case of saving 20~40 hours, when every day the child is usually found (manually) at Toy R Us, 15 minutes later. Are you telling me that there are surveillance operators at the mall constantly processing lost children tickets or scanning for shoplifters where this feature would be a big help? Or is this an application looking for a customer?

Robert, I think the more ideal hypothetical / aspirational use case is the London bombing (thousands of hours spent looking through video, etc.).

Toys R Us is not likely going to buy this but a big city with tens of millions of budget for surveillance, that I can see.

I am still not saying it will work well, but that's where the money likely lies.

All,

Not that I buy into this being ready for main stream but I can see the value. I have a bank robbery who has hit us at the same location three times. However after the second time the FBI wanted more then a months worth of video. they wanted to see if the guy had been to the location to case it out. If we did have some way of quickly sifting though the video to quickly pull up the body types this wuld have been helpful.

Anyway..... just my opinion.

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

Milestone Machine Learning Camera Auto-Setting Examined on Mar 13, 2019
Milestone wants to improve image quality using Machine Learning to solve the problem of "a camera doesn't know what it is being used for",...
AI Video Tester Released on Apr 02, 2019
IPVM has released the world's first AI video tester that lets you see how various AI models (including from Amazon, Google, Microsoft and YOLO)...
Covert Facial Recognition Using Axis and Amazon By NYTimes on May 20, 2019
What if you took a 33MP Axis camera covering one of the busiest parks in the US and ran Amazon Facial Recognition against it? That is what the...
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...
False Verkada 'Unrivaled' Low Light Performance Claim Removed on Jun 12, 2019
Verkada falsely claimed that it delivered 'UNRIVALED LOW LIGHT PERFORMANCE' until IPVM questioned. In fact, Verkada's low light performance is...
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...
RealNetworks SAFR Facial Recognition Profile on Sep 25, 2019
RealNetworks entered the surveillance market by giving away their analytics to schools for free, and is now targeting large commercial users with...
Gatekeeper Security Company Profile - Detecting Faces Inside Vehicles on Nov 14, 2019
Border security is a common discussion in mainstream US news and politics, as is the use of banned Chinese equipment by US Government agencies....
Ultinous European Analytics Startup Company Profile on Dec 04, 2019
European analytics-startup Ultinous pitches customers to "Have your own video analysis service!" We spoke to Ultinous to better understand their...
IPVM Opens 12,000 Sqft Testing Facility on Dec 16, 2019
IPVM is proud to announce the opening of the world's first video surveillance testing facility that will allow us to significantly expand our...

Most Recent Industry Reports

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...
Clearview AI Alarm - NY Times Report Says "Might End Privacy" on Jan 20, 2020
Over the weekend, the NY Times released a report titled "The Secretive Company That Might End Privacy as We Know It" about a company named...
Favorite Camera Manufacturers 2020 on Jan 20, 2020
The past 2 years of US bans and sanctions have shaken the video surveillance industry but what impact would this have on integrators' favorite...
"Severely Impacted" Mercury Security 2020 Leap Year Firmware Issue on Jan 17, 2020
One of the largest access controller manufacturers has a big problem: February 29th. Mercury Security, owned by HID, is alerting partners of the...
Apple Acquires XNOR.ai, Loss For The Industry on Jan 16, 2020
Apple has acquired XNOR.ai for $200 million, reports GeekWire. This is a loss for the video surveillance industry. XNOR.ai stunned the industry...
Installation Course January 2020 - Last Chance on Jan 16, 2020
Thursday, January 16th is your last chance to register for the Winter 2020 Video Surveillance Installation Course. This is a unique installation...
Halo Smart Vape Detector Tested on Jan 16, 2020
The Halo Smart Sensor claims to detect vaping, including popular brand Juul and even THC vapes. But how well does it work in real world...