Welcome to I-Cube: South Africa's leading provider of License Plate Recognition; Facial Recognition & Image Analysis   Products      News     Feed Back     Contact Us    Support      Where to Buy         
I-Cube advantage
I-Cube Biometric Selector
facts, features and benefits of Face recognition
Biometric Concepts
Face Recognition System Description
Facial ID
Card linked to Facial
System Design
       ID Verification
Value Prop.
Access control
Request Facial CD
Vertical market application (Casinos, Stadiums, Retail and Mines)
Proactive early warning crime prevention
HighVeld Steel Facial ID linked to TRACKING 2006
Business Park
Shopping Centres
Casino Access Control
CASINO I-CUBE Face Recognition Solution
Perimeter Solutions: Facial linked to Domes 
STADIUM I-Cube Face Recognition Solution
Selling biometrics to the retail sector
Where to Buy
STD Bank

Face Recognition user manual

Face Recognition FID user manual

Quick Start up guide

Installation & user manual 
FRS Discovery System OCX Control Reference Manual
Step by step technology guide to using face recognition
   Technical SPEC
   Start UP Guide
Solutions Roundup
Facial USES


Contact Us
\Feed Back
I-Cube Intro Brochure
What is FaceIt®

Local Feature Analysis
Technical Specifications

I-Cube Advantage

50 ways to use face recognition

CASINO I-CUBE Face Recognition Solution (PDF - 1.5 MB)

STADIUM I-Cube Face Recognition Solution (PDF 2.3 MB)

MINING I-Cube Face Recognition Solution (PDF 8 MB)

Selling biometrics to the retail sector (By BTT). (PDF 200KB)


I-Cube Face Recognition System design and costing


Design for existing cameras & digital recorder, laptop, R99 465.00       ON SPECIAL

I-Cube - Recognition VS Identity - the role of Biometrics (by BTT). (PDF 400 KB)

Face Recognition user manual

Economist comments on Face Recognition




 [ Request More Information ]

 [ Give Us Your Feedback ]



Subject: Looking for a better mouse trap

The comments below is based on a letter from: The Biometric Consortium's Discussion List 
Sent: 19 December 2003 12:42 PM

"Over the last 10 years, facial recognition has made considerable progress in improving its accuracy to verify the identity of enrolled users. Unfortunately, facial recognition has failed to recognize human faces as well as humans.

The speed and accuracy of biometric software and hardware presently available for fingerprint, hand geometry, voice, and iris recognition, and their ability to verifying and/or identifying a received image with a biometric identifier in the database, far surpasses the ability of individual persons or groups of persons to achieve the same level of accuracy within the same time span.

Conventional facial recognition as we know it today, has totally failed to keep up with the other biometric technologies. Humans can still verify other humans better than facial recognition software. The fact remains that each and every one of us knows at least 200 persons and each of these persons has a name that we use to address them. When we encounter one or more of these 200 plus persons, we are able to identify them from our human database within a second or two. We do not have a false identity level of 8% to 10% because they are standing in direct sun light, or under a fluorescent light in the office. Of course there are persons that are less known to us for reasons of a lower frequency of encountering these individuals on a regular basis. However, how often have we been able to identity an individual that we have not seen in over 10 years, has put on weight and gone bald? 

Perhaps, facial recognition has to re-think what is required in order to identify and/or verify an individual face. Is it the hardware or is it the software that needs to be replaced? Or is it both? Do we need to completely re-thinking what is required in order to be able to identify a face under various environmental conditions and facial expressions?

Fingerprint, hand geometry, and iris demand that a minimum amount of data be acquired in order to verify/identify an individual. A badly damaged fingerprint, a hand with a bandaged or damage finger, or an eye with a considerable amount of reflected light or eyelids covering the iris will always result in a less than acceptable rate to authenticate.

Can we therefore say that facial is no different than other biometrics? Unfortunately, in the case of facial recognition, this is not the case. It appears that in the real world the factors affecting facial recognition are more prone to alterations and occur far more often then they do in the biometrics mentioned. In addition, the other biometric identifiers cannot be modified by muscle movement or the re-doing ridges and minutiae, or altering the finger geometry, or changing the unique identifiers that surround the pupil. 

Facial biometrics is faced with various environmental factors that affect the ability of facial biometrics to accurately recognize an individual. The question is what can be done or is being done to resolve these issues?

The present development of several facial biometric vendors with 3D facial recognition systems is a definite step in the right direction. These vendors are still in the preliminary stages of product development but they have addressed the issues of facial positioning and environmental lighting as well as certain issues of facial expression."

I-Cube solves this problem but moving away from automatic recognition by computers, and using the best of computers and digital imageing and the human recognition abilities. "If we can take a moment and step back from the face and see the whole human being as a human being, maybe we can improve on an old but still relevant idea.

What happens when we confront a face and we are not exactly sure to whom that face belongs? How often do we say, That face is familiar, but I just cannot put my finger (no pun intended) on where I saw it and to whom it belongs." We know that we know him, but we cannot put a name to that face. We try placing that face into some historical incident that can tie that face to a name. Did we meet this face at a party, at the office, at school, or at a family gathering? We not only look at the face, but also at the age of the face, height, skin colour, and if we are lucky, the face may be in a context (at a blackjack table compared to a slot machine) that will assist us in knowing more about what is the face's possible identity.

Facial recognition has a lot going for it. It is user friendly, non-intrusive, easy to acquire, with a zero failure to enroll rate. 

Facial biometrics has been attempting to gain a major foothold into the physical access control market. We all know that over 99.99% of the persons accessing secure access points are legitimate users with access privileges. One of the major problems facing facial recognition has been false rejection of legitimate users. Only 0.01% of attempted access to secure facilities may be by an imposter attempting to access this area. What are the chances of such an imposter succeeding in accessing this area if we were to layer both facial and other access control methods (a pin and a card)? What are the chances of denying access to an authorized person if we combine both facial and other access control methods? By combining biometrics with normal access control methods a higher level of security can be obtained. 

My position is not an emphatic YES for combined facial and normal access control methods, but rather a re-thinking of our one-track thinking that follows the line that biometrics must remain independent of standard access control systems and remain true unto itself.

Biometrics need to be far more open and recognize that it can learn from access control. By the biometric community and access control companies working together, we just may be able to make a better mouse trap. 

The comments above are based on a letter from: The Biometric Consortium's Discussion List 
Sent: 19 December 2003 12:42 PM
Subject: Looking for a better mouse trap
LISTSERV members may access the BIOMETRICS mailing list archives or change their subscription settings at: http://peach.ease.lsoft.com/archives/biometrics.html.


US airports use biometric checks

US airports use biometric checks

Millions of passengers leaving United States' airports face mandatory fingerprinting under new security guidelines, The Guardian newspaper reports.

Passengers who travel to and from the US will have to present their index fingers - as well as their passports - at check-in, from the end of next year, according to a senior security official.

Michael Jackson, the deputy secretary of the department of homeland security, said the procedure would apply to all passengers and airlines flying out of the US, as the country accumulates a mass of information on every person travelling through America. "What we are trying to say is that it's not enough to give biographical data. We will need biometric as well as biographical data."

UK schools fingerprint pupils

A survey of local education authorities (LEAs) discovered 285 schools regularly fingerprint pupils and store their biometric details on record, adding the real figure could be higher, reports politics.co.uk.

Despite this, the Department for Education and Skills has not issued any guidance on when and how biometric data should be collected and stored. Only a quarter of LEAs have any guidance available and in the vast majority of cases do not know if parental consent was given to collect fingerprints.

A Liberal Democrats education spokeswoman, Sarah Teather, said: "These figures confirm an extremely worrying situation where schools are fingerprinting pupils without any guidance on whether it is legal to do so.

Savi joins RFID project

Shanghai International Port Group (SIPG), the Georgia Ports Authority and Savi Networks announced plans to launch an RFID-based network that automatically tracks the location and security of containerised cargo transported between the Port of Shanghai in China and the Port of Savannah in Georgia, CargoNewsAsia reports.

Called the “Shanghai-Savannah Express Trade Lane Project”, the companies announced their partnership plans at the third annual China Trade and Logistics Conference held at the Savannah International Trade and Convention Centre.

SIPG and Savi Networks have jointly developed a localised China RFID solution that includes electronic seal and GPS integrated handheld for this project. The solution will be built upon an open, international standards-based network platform.

Data volumes drive RFID

The European RFID market is set to achieve significant progress, given its past and existing initiatives in retail, healthcare, manufacturing, logistics and anti-counterfeiting measures, as well as the increasing investments in this space, MoreRFID reports.

While this market has not demonstrated growth to the same extent as the RFID market in the US, it is nevertheless poised for growth. New analysis from Frost & Sullivan finds that the markets earned revenue of $41 million in 2006, and estimates this will reach $181.8 million in 2013.

There is a strong case for RFID in places where the unit cost is high, and also in places where inventory loss is widespread. Likely growth segments for RFID middleware will be in the areas of drug manufacturing and tracking, medical equipment tracking in hospitals and asset tracking.

Oz delays smart card introduction

Australia's official opposition party says their government has delayed introducing its smart card policy because of community opposition, ABC Newsonline reports. The government announced the health and social services access card may not be introduced until next year.

Labour's human services spokeswoman Tanya Plibersek has told the Australian Smart Card Summit in Sydney that the Government-proposed access card is too expensive and there are still privacy concerns.

"What happened yesterday was a bit of excess baggage that we've ditched from the saddle bags on the way to an election," she said. "I think this proposal has been jettisoned because it's been publicly unpopular, not because of the technology, but because of the whole framework that has been presented to the public."






           Call +27 (0) 82 562 8225

I-Cube.   All rights reserved.  Revised: February 18, 2008 .






I-Cube.   All rights reserved.  Revised: February 18, 2008 .