D.M. Nemtinov, A.V. Mishchuk, O.F. Antonov
Problem statement. Biometric technology is being intensively implemented in all areas of business at the present time including passport control processes to identify a person at border crossings. The conditions of image registration during passport control significantly affect the probability of successful verification as different biometric algorithms respond differently to interference factors. The conditions of registration differ depending on the checkpoint, so it is difficult to implement a universal algorithm for verification. The proposed solutions for verification of persons on open sets do not always meet the requirements. However, this is largely due to insufficient development of the methodology of the experiments. As a result, it is not possible to quantitatively compare several biometric algorithms to unambiguously select the best one for a particular application.
Objective. To develop a methodology for comparing biometric algorithms by quantitative characteristics to solve verification problems in passport control.
Results. A methodology for comparing biometric algorithms based on their probabilistic characteristics obtained in accordance with international standards has been developed.
Practical significance. The resulting methodology can be used at the stage of operational tests of automated systems that solve the problem of identity verification by biometric features.
Nemtinov D.M., Mishchuk A.V., Antonov O.F. Method of comparison of biometric algorithms in software and hardware complexes. Radiotekhnika. 2021. V. 85. № 2. P. 64−68. DOI: 10.18127/j00338486-202102-10 (In Russian).
- Kuharev G.A. Metody obrabotki i raspoznavanija izobrazhenij lic v zadachah biometrii. SPb: Politehnika. 2013. 388 s. (In Russian).
- GOST R ISO/MJeK 19795–1–2007. Avtomaticheskaja identifikacija. Jekspluatacionnye ispytanija i protokoly ispytanij v biometrii. Chast' 1. Principy i struktura. M.: Standartinform. 2009. 52 s. (In Russian).