Integrated Biometric Recognition System of Person Identification

V.G.Yakhno1, E.V.Eremin1, A.A.Tel'nykh1, M.A.Kostin1, A.N.Kogan1, A.E.Ivanov1, J.Kh.Radzhabova1, N.S. Bellustin1,A.O.Perminov1, V.A.Ruzumov1, D.A.Tikhomirov1, S.G Shilin1, M.A.Sorokin1, U.S.Rohatgi2, and A.K. Jain3, 1 - Institute of Applied Physics RAS, Russia, 2 - Brookhaven National Laboratory, USA, 3- Excom,Inc, USA

Technology of Integrated Biometric Recognition System - IBRS is developed. Images from deferent biometric sensors (Hand, Face, Fingerprint, and Voice) are used for features calculations. Algorithms are developed in accordance with BioAPI specification for Hand, Face, Fingerprint, and Voice. Use of these technologies in any combinations.




Differentiators

1. Standards Based Implementation
   a. Biometric industry standards APIs
   b. Industry popular databases - MS SQL and Oracle
   c. Industry standard encryption for template transmission & storage
   d. Data storage per NIST standards
2. Most Accurate System available in the market
   a. Uses Neural Network Logic to adapt to screening conditions and users profile
   b. Layers and Integrated approach increases accuracy by 10 folds
3. Easy to Use and Flexibility
   a. Integrated (more types of scans) and Customizable Applications
   b. Runs on Commercially available HW/SW platforms (Cheaper to deploy)

The demonstration version of IBRS is a complete recognition system with biometric-data input devices, a system for control of parameters and recognition scenarios, separate multiple client stations, and the central recognition server. The server stores biometric characteristics of users and performs input-data recognition operations. Confidentiality of transferred data is ensured by the use of modern digitization technologies. The IBRS is operable with the use of commercially available devices that ensure recording of biometric data. For distributed calculation within the framework of the IBRS, we developed a module CosBioAPI (CORBA Service for BioAPI) for provision of network transparency for using the AwpBioAPI recognition kernel. Moreover, by using the Oracle database on the central recognition server and the recognition algorithms scalability, the possibility of system operation with a large database (up to 1 million users) is demonstrated.

Consideration of features from different multiple sensors and adaptive algorithms allow decreasing the error level by more than tenfold in our Integrated Biometric Recognition System.

Possible application:
Identity management in fraud control; Ease in E-commerce; Better customer service; Control of passengers during boarding at airport; Immigration control; Security.

Principal Investigator of IPP - 304 - BNL project: Vladimir G. Yakhno (yakhno@appl.sci-nnov.ru) - Doctor of Sciences (Physics&Mathematics), Head of AWP Laboratory Institute of Applied Physics of RAS.

Download: