Some Tech Comparisons / Study

Thought would just start a separate thread on this so not lost in the general discussions - doesn't have to be carried on but thought some who like to look at the mechanics / studies / projects around various biometric tech comparisons may be interested.

This was a study / project commissioned / supported by the EU 7th Framework Programme into various biometrics and their capabilities, adv / disadv. etc. for Mobile Phone implementation, started in 2013 and handed down in Jan 2014 (so a little aged but still relevant data IMO). Just must remember that advances in the tech would have been made in the last 12 mths.

Warning - it is long but relevant sections can be skimmed to....lol

One of the key general consensus I have found in my research is the EER figure as being one of the best measurements of a biometric product capabilities although most commercial entities don't advertise that # preferring to use best mkting #'s like FRR, FAR etc. that sound better (there are articles out there about this type of mkting #'s). That, and what is the least invasive / easiest to use consumer wise.

The EER (or sometimes called CER) is the point where the FAR & FRR cross or are equal and therefore by having a lower EER / CER you have best case scenario for FAR & FRR. There is always a compromise on FRR & FAR (one up, one down etc.) to achieve EER so you can see why some Co.'s only "advertise" certain numbers but not the "real" EER.

Right, now that is said, I thought I would post this so people can skim through to get maybe a better understanding of essentially our competitors in a purely objective manner.

The key to look at is the various modalities and their corresponding EER's. Iris from what I quickly skimmed through appears best performing with low EER's.

I started looking for EER comparisons as I noticed a while ago that the WBT site actually state their EER at 0.85% - purely on the corneal biometric.
"Milestone

Live trials hosted at University of Tennessee on many hundreds of subjects and three iterations of the technology returned a low EER of 0.85%"

What must be remembered is this rate was achieved quite early on in the timeline as it was at the initial Uni of Tennessee testing - around a yr or so ago + and WBT have refined and had many new iterations of their product and enhanced further with multi-biometric performance now including Iris as well and also state on their website their commitment to R&D and driving this number down further through 2015.

Fusion or multi-modal / multi-biometric is the next wave as being able to combine various methods enhances security but it is a fine line having the end user overloaded with having to complete various things just to unlock a ph. You know - rub your head, pat your stomach, spin 3 times etc.....lol

WBT appear to have found a nice balance IMO as per below statement from the website:

"Wavefront works specifically to incorporate liveness and captures multi-modal biometric information in one pass. All this in close to real time. Simply, quickly, with low user effort."

All this achieved in visible light. In comparison with the project study (iris section) which utilised NIR/IR to achieve their results with one of the disadvantages listed being specular reflections which just happens to be one of WBT's strengths

Is an interesting skim read for an independent view IMO if you wish. Enjoy - off to do some real work now

https://www.pcas-project.eu/images/Deliverables/PCAS-D3.1.pdf