AGNITO CEO discusses innovations in voice biometrics and 2016 strategy
AGNITIO has long been a provider of voice biometrics solutions for government and commercial sectors.
The company’s series of KIVOX voice biometrics products are integrated into the security solutions of financial services, telecommunications, utilities, and companies in other sectors.
In addition, AGNITIO’s underlying voice ID technology is deployed in the most demanding environments, such as forensics and intelligence.
For example, Morpho recently integrated AGNITIO’s voice biometrics engine into its multi-biometric platform, MorphoBIS, to provide law enforcement and intelligence agencies with a set of advanced multi-biometric tools for identifying suspects and leads during investigations.
BiometricUpdate.com had the opportunity to talk to Emilio Martinez, AGNITIO’s CEO, where he discussed the company’s sale of mobile assests to Cirrus Logic, partnership with Morpho (Safran) to provide voice biometric solutions for law enforcement agencies, and AGNITIO’s voice biometric solutions innovation and overall strategy for 2016.
What is the new focus of AGNITIO after selling off its mobile assets to Cirrus Logic?
Emilio Martinez: AGNITIO continues to offer the best of breed server-based voice ID technology. Voice biometrics technology has been used and will be used in multiple server based applications. We are improving the speed and accuracy for critical applications of our government customers. And at the same time, providing an accurate and easy to use voice biometric engines for financial services customers. fraud detection, proof of life for pension payments, two-factor authentication and vocal signature are the main applications. There is a huge market where server based biometrics will be growing and we continue to provide the leading technology there.
I understand that AGNITIO partnered with Morpho (Safran) to provide voice biometric solutions for law enforcement agencies to identify criminals and protect citizens. Can you describe in more detail which products have stemmed from this partnership?
Morpho is a global provider of biometric technologies — from voiceprints to DNA and face. AGNITIO was selected as the supplier of voice biometrics to complete their offering. We already had many common customers in the government space and we are now in the process of integrating our technology in most of their multimodal solutions. The first example was already presented at the Biometric Consortium conference in Tampa: the MORPHO BIS product. The USA branch of Morpho, MorphoTrust, also has a very powerful cloud-based multimodal biometrics platform called Identix aimed at the government and enterprise markets. Our technology is being integrated into this solution as well.
At the same time we are also providing text dependent technology to financial services customers of Morpho. When Morpho acquired Dictao in 2014, we were already working with them in some PoC’s for some big banks in Europe. We continue this path and there are several projects in progress where KIVOX 360 is used as the authentication engine.
Why do voice biometric solutions provide an effective identification method for organizations and government agencies?
The main use of voice biometrics in government agencies and organizations is mainly to identify criminals, verify identity and detect fraudsters in call centers using voiceprint black lists. Active and passive authentication is just a small portion of the use cases for voice biometrics. Voice ID is also the authentication factor of choice for multiple applications and for many reasons.
We strongly believe that future authentication is in the multi-factor domain — not only because of security (fusion of scores of independent factors can provide a more accurate decision), but because of convenience. Authentication will happen in a variety of situations and with a large choices of devices, and no factor will be ideal for all situations. If you are driving or walking in the street using your hands free or mobile phone, typing is not a good option so voice can be perfect, and can be combined with other factors such as mobile or car metadata, position, etc. When you’re in a meeting or conference you will not want to say something in a quiet room to open your recent mail. We are also sure that voice will always be one of the factors that users will like to have as a choice and vendors will prefer for their multi-factor solutions.
Users will prefer because is really friendly and generates less privacy concerns that other factors. Your voice is there and you share with others continuously. And vendors will prefer because it is really cheap to implement; you already have microphones in almost all devices that people are using to authenticate. There is no need for extra HW (sensors, special cameras) to get the voice and transmit to a server. What you need to complete this picture is to provide an accurate technology, which is protected against very obvious spoofing techniques such as recording your voice and replaying it for authentication. AGNITIO’s engine, after more than 10 years of experience in the field, is able to provide that robust technology. And the combination of ease of use, cheap implementation and secure results is the recipe for success of voice as a relevant player in this multimodal authentication space.
What are AGNITIO’s plans for 2016 in terms of voice biometric solutions innovation and overall strategy?
2016 will be the year for a new technology generation. Our fourth generation technology (4G) based on advanced i-vectors algorithms was created in 2012, and now all of our products are using it at full potential. At that time It was absolutely disruptive in terms of speed and accuracy, enabling the use of voice biometrics in new spaces. Our government customers were able to obtain results that were impossible until that moment. And our commercial partners were able to implement very successful free speech fraud detection and authentication products in the call center.
We work hard to maintain leadership in the voice biometrics engine market for free speech applications, and in 2016 we plan to launch a new generation including advanced neural networks algorithms, better accuracy in noisy environments and 16Khz capabilities among other new characteristics. It will boost the performance and dramatically increase the accuracy in more realistic environments. The 5G version of KIVOX 360 (our text dependent product for collaborative authentication) will also take advantage of these improvements allowing a more accurate and robust enrolment and authentication process in all channels: call center, web and mobile using a single server based engine and a single voiceprint.