Interview with Martin George, CEO, Smart Sensors Ltd.

 

altFB
Can you please provide our readers with a brief background of the company?

Smart Sensors
Don Monro was Professor of Information Systems at the University of Bath (UK), and was running the Signal and Image Processing Research Group there. I met up with him initially to discuss how some of his innovative work on video encoding and compression might translate into use with CCTV surveillance and security systems, which were becoming a very hot topic in the UK in the early 2000’s. I already had a background in biometrics, and we quickly got to talking about opportunities in that space. So between us we sponsored the development of various ideas via research projects under his guidance at the University. Out of this work not only came proof that we had something new and valuable in iris biometrics, but also the University of Bath Iris Image Database which has become a valuable resource for many other members of the research community in iris biometrics. Smart Sensors Ltd. became the commercial vehicle through which this research started to get exploited in 2006, and one of our first contracts was to prove that alternative algorithms could work efficiently with the Sarnoff “Iris On the Move” system, then in its early days.

FB
Your company was recently awarded the Frost & Sullivan “2009 Global Iris Biometric Systems Technology Innovation of the Year Award for your efforts in developing the Monro Iris Recognition Library and INterface Software Development Kit. Can you please tell us more about this award?

Smart Sensors
Yes, we were excited, surprised and honoured to get this award, and I think it reflects the hard work our team has put into our research and development programme over the past couple of years. All of the algorithm developers, ourselves included, have initially focused on feature extraction methods that code the spatial frequency features of the iris texture, once segmented from the image of the eye. The innovation that I believe swung it for us was our recent work in encoding the iris image by vector methods. These give results that are different from spatial frequency methods, and by data fusion of the two methods, can lead to exceptionally low error performance. The vector methods also show great promise with partial iris images – one of the key goals of current research programmes.
Another aspect has been the relatively small computing footprint that our technology requires, and the fact that it can exploit existing techniques that are themselves already well-proven in embedded hardware designs. This should lead to a good future in low cost high volume products.

FB
Can you please review your company’s offerings?

Smart Sensors
Our main offering is the MIRLIN SDK (Software Development Kit), now in version 2.11. This is a fully featured set of development tools primarily for Windows XP/Vista/Server platforms, and for use with Visual Studio or other C-based development toolchains.
Our philosophy has been to make this SDK useful to developers not only in building products, such as cameras, and systems in which iris biometrics are to be integrated, but also for researchers who want additional low level functions and diagnostic data exposed, such as a variety of image quality metrics, boundary and centre-point co-ordinates for pupil and iris, as well as metrics about features such as specularities and so on which are different according to which manufacturer’s camera you use, and often, the environmental conditions too. Recently we have added much more user control over how templates can be used: especially how matching is scored, and how eye rotation can be handled, including proprietary methods of dealing with less robust parts of the iris data.

We have also built versions of the MIRLIN SDK for LINUX, Windows CE, Windows Mobile, and TI OMAP family DSPs.

Our other offerings include a fast iris/pupil finder for use with cameras, and a tool we call MIRLIN Iris Analyst, which permits a comprehensive analysis of error and metrics performance on a target database of iris images. We also offer the University of Bath Iris Image database, but for research and evaluation purposes only.

Finally, “Vector-MATCH” is our new capability in vector analysis and matching of iris images. Although this is not available as part of our SDK, the technology can be licensed from us.


FB

Please review a recent deployment for us.

Smart Sensors
A good example would be our work with Datastrip’s DSV family of mobile multi-biometric and credential verification units. We built an iris recognition engine for Datastrip’s DSV II mobile unit, which runs Windows CE on a small footprint processor, with 8 hours continuous battery life. The unit can read virtually every type of mag stripe, chip card and contactless credential, including barcodes, and can match in excess of 4000 records on-board with iris, face and fingerprint biometrics in addition to a full set of biographics. Greater matching capacity can be had via built-in wireless network communications.

Virtually the same development also works on their new DSV III family, which employs a more powerful x86 family CPU with consequently greater capacity.

The Datastrip units are in use by a variety of US service agencies including the military and the Coast Guard, as well as in other deployments worldwide including various police services.

FB
Has this past year been a good one for Smart Sensors?

Smart Sensors
Yes. I believe we’ve come a long way in our technical development, and we’ve secured a number of major licences and partnerships which we hope are going to bear fruit both for us and our partners in the coming years. We’ve also tried to give our partners very good commercial reasons for wanting to work with us. Smart Sensors is seen as an algorithm provider, and although I think our relationship with clients goes well beyond this into the technology and human factors expertise that lie behind making iris work as an efficient and strong biometric, nevertheless for a small company like us the headline perception may remain. I hope we can convince new customers that there is much more to us than just algorithms!

FB
Where do you see the Iris Biometric market moving to over the next 5 years?

Smart Sensors
The market for Iris Biometrics is, compared to fingerprint and facial recognition, still immature, and very much geared to the USA – with certain exceptions in Asia. But that’s changing as new competition comes along. We’ve also been one of those companies at the forefront of trying to change the commercial model to one that could facilitate the mass deployment of iris biometrics at very economic prices. There is the prize of mass deployments for a variety of what we call “entitlement” opportunities in emerging markets where the environmental and demographic conditions make iris a firm contender over fingerprints as a strong biometric. Then again the US government has said that it supports the idea of non-contact biometrics at border entry and exit points. So that will be good news over the next few years for the development of Iris on the Move and Iris at a Distance technologies such as are embodied by AOptix and Sarnoff. Getting the “human factors” right is critical to making this concept work, and maximising throughput.

Speaking for my company as a foreign entity operating outside the USA, I have to acknowledge the part played by the US government in its willingness to sponsor and experiment with R&D activities that have been instrumental in driving many of these developments forward.


FB

What can we expect to see from Smart Sensors in the near future?

Smart Sensors
As deployment costs come down and usability improves, the non-contact and convenience benefits of iris biometrics are going to result in an increasing share of deployments. We have some exciting work going on in the field of combining logical and physical access control systems, which we will be talking more about in the London Biometrics 2009 conference, but in essence it puts another cornerstone into the rapidly emerging market for smart buildings and site management systems. It makes a lot of sense for integrated buildings systems to take care of security as well as energy management, as both are linked to who is currently in the building, or on site.      www.smartsensors.co.uk/

We’re also still developing our intellectual property portfolio, so you can expect to see a number of innovations move from research into our product offering over the coming year or so.

Finally, and most importantly, I look forward to continuing working with and developing our partners who are of course crucial to getting innovative and competitive products and systems out into the market. They themselves also have an exciting story to tell.