Interview with Li Wang, CEO, Warwick Warp Ltd.
FB: Would you please provide our readers with a brief background of the company and its relationship to the University of Warwick?
Warwick Warp (WW): Warwick Warp is a spin-off company from the University of Warwick Image Processing Group and the Department of Computer Sciences. The founding members, including myself, all have strong image processing backgrounds, image analysis and pattern recognition. When we set up the company, we assessed the market and concluded that there was a technical gap for fingerprint biometrics that dealt accurately with low quality data.
When the fingerprint or other biometric data quality is high current technology is able to accurately detect those cases, problems occur when the data is of a low quality. The modelling technique we have created provided a positive impact and increased the overall system format, especially when the fingerprint quality is poor. So, on that basis, we started Warwick Warp, developing a mathematical modelling technology to retrieve underlying ridge structure from noisy data.
Underpinned by our mathematical modelling technology, we began to develop a number of software products to address the civil and law enforcement AFIS market, as well as access control, registration and financial transaction applications, whether that be physical or logical access control.
FB: Therefore, your technology approaches image recognition in a brand new way; could you describe some of the features of your approach?
Warwick Warp (WW): Certainly, the conventional technique, largely utilizes image filtering methods that rely completely on the input data. When the data quality is high, then the technology seems to work well. When the data quality is low, it tends to be sub-optimal. A trained human, would be able to recognize the underlying ridge structure very well, even when there is strong interference from the background or any other noise in the data. That is because a preconceived model of what the fingerprint should look like already exists. Therefore our technology is designed to imitate the human approach thereby using a mathematical model to retrieve the underlying structure and not solely rely on the input data that is presented to it. It has a model, a prior knowledge of what it is looking for and can therefore accurately recover the underlying ridge structure from low quality images. That is really the core uniqueness of our technology.
FB: Now, you’ve been recognized by NIST, (National Institute for Science and Technology), and I believe that you participated in the proprietary fingerprint technology test in 2010; would you please share the results of that testing with us?
Warwick Warp (WW): Absolutely; we are very pleased with the test results. The test indicates that the performance of our technology is, actually, achieving, if not exceeding, the current state-of-the-art technology. We are a fledgling company that has created a brand new technology without massive resources and time at our disposal and have managed, on the first and second submissions, with very, very high accuracy compared against market leaders, to create an accurate and fast fingerprint technology. So, yes we’re very proud of our NIST test results. Furthermore, as this is a brand new technology at its early stage of the growth curve, we are expecting substantial further accuracy improvement in the near future.
FB: Your company recently entered into a partnership with WCC, out of The Netherlands, would you please elaborate on that for us?
Warwick Warp (WW): WCC & Warwick Warp, are really two companies that strive to provide high quality products at a competitive price. WCC provides an excellent platform, a multi-mobile fusion platform, to utilize the strength of different modalities of biometrics. Warwick Warp, as a core AFIS provider, will provide the core fingerprint identification functionalities for the WCC ELISE product. So, the partnership will bring two great products with great quality together and offer a competitive solution to system integrators or end-users.
FB: Am I to understand that WCC approaches the marketplace in a similar manner to that of Warwick Warp, in that they use mathematical models to try and determine results?
Warwick Warp (WW): Well yes, you may say that, but WCC’s business model, and product is very different from Warwick Warp. WCC’s business base is a platform; WCC’s product doesn’t itself contain any core matching capabilities in terms of fingerprint, face or iris, it is a management middleware suite and the solution to actually combine the results that were generated, for example, from our software and that of other face or iris matching results.
FB: So, you could say it is like a perfect fit then?
Warwick Warp (WW): Yes, absolutely.
FB: Could you please review your product line-up for us, including the recently launched Synthesis Simplicity testing tool kit?
Warwick Warp (WW): Sure; we have, at the moment, four main products. LiveScan SDK, which deals with small to medium scale lifescan deployments, and matching and template creation. The second product is Warwick AFIS SDK, which is designed to deal with large scale applications, one in which the number of users exceeds several hundred thousand. The third product is Warwick Warp latent AFIS, which is specifically designed to deal with a law enforcement application, and it has the capability to match, with or without human intervention, the print to mark, mark to print, or mark to mark comparisons. The last product is the one we have just released, which is called Synthesis Simplicity, and is a tool designed to test biometric systems. When we talked to systems integrators and end-users, we found that in order for them to test, or evaluate, the performance of their systems, a large number of biometric or fingerprint data, would be required. This is because of the nature and sensitivity of the real fingerprint data, as large scale databases test data is very difficult to come by. The Synthesis Simplicity tool really addresses this shortfall of testing “synthetic” data. It has the capability to generate realistic fingerprints and it the capability to integrate and evaluate realistically, using the data that is generated by the tool, which can be on a very large scale, without encountering data protection issues that are associated with real data.
FB: What is next for your company? Are you going to apply your innovative technology to other modalities, such as facial recognition, iris and etc.?
Warwick Warp (WW): Well, we have a number of products that are related to fingerprint, and we would like to increase the features that would fall within those products to provide faster, more accurate and more user-friendly products associated with the fingerprint market.
FB: What do you see as the challenges that are currently facing the industry; especially as it related to activities in the UK?
Warwick Warp (WW): I would say that the high-end biometric market –used to, or has been, heavily relying on government procurements; with the current financial situation of the UK government and internationally, the market is going to be tough. However, my view is that this will be more than offset by the strong growth from the commercial sectors and other emerging markets. What markets want now and in the future is a biometric solution that is robust, accurate, fast and easy to deploy. The market should not be constrained only by the governments’ applications. There are many other applications that would require accurate identity management through biometric technologies with a high degree of accuracy and speed. Warwick Warp is in a unique position to provide such solutions.
FB: Do you see things, like the mobile identity marketplace, starting to grow for financial payments, etc.,?
Warwick Warp (WW): Oh absolutely; obviously, the infrastructure needs to be put in place first; but providing the technology could achieve the level of accuracy and speed that is required in those applications, then it is only natural to use biometrics as a means of identity verification. Obviously, within the mobile market, whether using within financial transactions, or another purposes, because the nature of the application intends to provide a larger percentage of low quality data, as opposed to a fixed device. There, our technology would be able to provide a useful solution due to its capability of dealing with low quality data more accurately.
FB: Yes, and not just for mobile, but for any purpose, for that reason really, your technology would be advantageous.
Warwick Warp (WW): Yes, absolutely; in situations such as when dealing with military contexts, due to the hostile environment, it tends to produce a larger number of low quality data. Or if certain country or demographic area have a larger percentage of manual labourers in the population, then our technology would be able to provide a more accurate solution than many of our competitors
FB: Thanks you very much for speaking with us today.
Warwick Warp (WW): It has been my pleasure Peter.
Warwick Warp Ltd, a spinout company from the University of Warwick, has developed a novel fingerprint recognition technology based on the shortcomings of current systems such as their inability to deal with poor quality fingerprints.
Warwick Warp’s algorithms overcome these shortcomings by applying an adaptive mathematical model to model the ridge structures as well as allowing for both linear and non-linear deformation in the measurement. Furthermore, the technology can robustly detect and incorporate what are known as third level features into the measurement such as sweat pores which have the potential to further increase accuracy in the match.