Interview with Tore Etholm-Idsøe, CEO, NEXT Biometrics
Recently, NEXT Biometrics initiated a study from the University Carlos III of Madrid examining the correlation between fingerprint sensor size and accuracy. The results showed a positive correlation between the two factors: a larger sensor increases the quality of fingerprint identification.
FindBiometrics President Peter O’Neill had a chance to speak with Tore Etholm-Idsøe, CEO of NEXT Biometrics, on these study results as well as the company’s success in the wake of iPhone Touch ID technology – a phenomenon NEXT refers to as The Apple Effect.
Peter O’Neill, President, FindBiometrics (FB): We have heard there is an extensive university study regarding sensor size and quality recently completed. Can you give our readers the highlights of this study?
Tore Etholm-Idsøe, NEXT Biometrics, CEO (NEXT): At the University Carlos III of Madrid they conducted the first comparative mass market realistic fingerprint sensor study of its kind – this to prove the performance of NEXT Active Thermal Principle and also to demonstrate the link between sensor size and quality in fingerprint identification. The study, just released, will raise eyebrows in the industry because the results are both very clear and very important. When you decrease the area of a fingerprint sensor there is a dramatic increase in the number of false rejections. This confirms what biometric experts have claimed all along – that size is the most important determinant of a sensor quality system and you cannot use small sensors without sacrificing quality.
It has within our industry been a common practice to claim heavily exaggerated performance numbers. Even for the swipe sensors, which are now finally almost discontinued, false rejection rates under 1 percent were claimed at false acceptance of 1 in 1000. Anyone that has ever been in contact with a swipe sensor notebook or smartphone knows that this is very far from the truth. We have met major manufacturers of notebooks that have stated that more than 95 percent of their swipe sensor customers chose not to use these sensors. This is due to sky-high false rejection rates. Going forward, our industry needs to mature. However, now we again see a tendency of severe sensor shrinkage accompanied by claims that “we fix the lack of area in our algorithms.” One can simply not make such claims without backing this with real mass market, credible 3rd party comparative tests.
The “Madrid Study” initiated by NEXT, is an independent ISO standard best practices methodology study done under the management of Professor Raul Sanchez-Reillo, a highly respected resource in the biometric world. What they did was to compare two market leading capacitive area sensors and our sensor. The origin of the sensors used was kept confidential to the operators, the sequence of sensors for print collection was randomized for every test subject, the population was representative in a mass market context and all test subjects were also requested to do the acquisition in 2 sessions, with minimum 2 weeks break between each test.
The study proves two things. First, it proves that our Active Thermal Principle and the other same size capacitive high cost sensors perform similarly. Secondly, the study proves for all sensors that there is a direct link between the size and the performance of the sensors. Obviously the subjects used had different ages, they had different quality fingers, some had dry fingers, some had wet fingers, some had dirty fingers, some had cut fingers, some had worn fingers, and as such the subjects were representative as a mass market population. Now what the study also did, was that it cropped areas of the prints collected to simulate what would happen if the area was smaller. This was done at 10 x 10 mm and also at 8 x 8 mm.
The study very clearly demonstrates that when you decrease the area of the fingerprint sensors there is a dramatic increase in the number of false rejections. This is something that a lot of individuals around the world experience every day. You simply cannot use small sensors in a context of population scale use. Sure, it will work well for many individuals but if you sell systems where the sensor system is dependent on working for a high percentage of a given population, for example with employees of an organization or customers of a bank, you will have a major problem with a significant percentage of the users unless the sensor sizes are big enough to give the best accuracy and reliability and avoid false rejections.
Also bear in mind that the Madrid study has been performed in an in-door environment. In an out-door use case (e.g. smartphones) with impatient or careless users, the frequency of wet fingers, dirty fingers and skewed placements will increase. In such use cases the dependency of sensor size is even more critical – the false rejections rates of small sized sensors rises to levels found unacceptable by most users.
FB: So what are the implications of these findings for NEXT as a company?
NEXT: The advantages highlighted in the study relates to the core, exclusive and patented technology. The fact that we are not using a silicon area sensor, but a polysilicon production process, gives us a huge cost advantage compared to all capacitive sensors. Even compared with tiny area, low performance competitors there is a significant cost difference. Today, NEXT is the only player able to deliver mass market volumes of high quality sensors at prices that are attractive.
To put it in other words, our competitors need to cut sensor size in half or even a third to reach a price that is acceptable for high volume applications. In doing that, they sacrifice quality, a fact that is often clouded by talk about algorithms and other “fixes” to the quality issue. But, there are no fixes for this by nature of the application sensor size matters and it matters the most.
FB: Can the fingerprint industry expect turbulence and problems regarding the IP situation, given the present number of suppliers in the smartphone space?
NEXT: Absolutely. The capacitive sensor market is a congested space with a lot of patents close or overlapping. By far, Apple has the strongest portfolio. With its own and exclusive core principle, NEXT has a very clean IP situation, a factor which we believe an increasing number of high-volume customers will appreciate.
FB: Are you ready for mass production today?
NEXT: Yes NEXT has been in mass production for a while now. Further ramp-up activities are on track and from May, 1.5 to 2 million units per month is feasible and fairly easily scalable.
FB: Now mobility is a key driver in our industry, are you expecting significant growth at NEXT as usage grows in mobile and other devices?
NEXT: Yes, we see a significant “Apple effect” not only in the mobile sector but also in tablets, notebooks, traditional markets and in what we call “NEXT-enabled” markets.
We are definitely very competitive in notebooks and tablets and also among smartphone players that want to differentiate on quality. NEXT sensor systems are compatible with the needs of payments, corporate access and other coming security dependent applications.
We also see a fast increasing number of requests in a long range of new marketplaces. These markets are typically small, often handheld devices that need to be manufactured for $10 to $15 USD. But there is a challenge with these applications. Opposite to smartphones, there is no way to enter a pin-code or password if the sensor fails. So, in other words, these markets demand a high quality sensor system at a very low cost. We have labelled these markets as “NEXT-Enabled,” because they are basically made possible by our unique combination of quality size and low cost. We see a lot of interest in this segment. Just this past January we got six sample orders from companies in that market segment. Importantly, a couple of them were high quality brand name players that have seen what is happening in the mobile space and now want to see if they can fit this type of user authentication into their own devices.
In sum, given our present products and 2015 roadmap, NEXT has prime competitiveness access to larger markets than any other player in the fingerprint sensor industry today.
FB: How are you working to gain access to the Tier 1 and Tier 2 customers?
NEXT: We have a combination of direct and partner-based sales. Our direct sales organization has grown very strong. Experienced biometric and IT industry people easily recognize our competitiveness and they join us. Most recently, this was proven by the fact that the former President of PC Global Operations at ACER, Campbell Kan, joined us to take care of NEXT Operations and Asia Sales. He is obviously very well connected in the Asian notebook, tablet and smartphone space.
High volume sales processes take a long time. The market obviously knows we secured a multinational brand name Tier 1 contract in November of last year. This customer will start launching products in the third quarter of this year. Before signature we had already worked with them for twelve months and we are now two-thirds through an 8-month process of joint preparations. We have a number of such sales processes going on and expect to announce more news on this in the upcoming quarters.
FB: Well Tore, it sounds like very exciting times for your company. Thanks for sharing the findings from the Madrid report study.
NEXT: It has been my pleasure Peter.