V.A. Ivanov, A.A. Smirnov, D.A. Nikolaev

Statement of the problem. A significant place in biometric identification systems is given to face recognition by images. This is a complex statistical task, which is often carried out using artificial neural networks. The process is based on a comparison of a reference photo and a test one. At the same time, it is necessary to take into account a number of destabilizing factors – lighting, shooting angle, aging of the object, etc. The greatest interest from the point of view of biometric identification for security systems is the assessment of the quality of face detection depending on the time elapsed since the receipt of the reference photo.

Goal. To investigate the characteristics of detecting the desired faces of people in images obtained using artificial neural networks, taking into account the time elapsed since the receipt of the reference photo.

Results. A method for evaluating the detection ability of artificial neural networks in the conditions of aging of the desired object (human face) has been developed and tested.

Practical significance. For the systems of biometric identification of persons, the dependence of the deterioration of the quality of the work of artificial neural networks in the conditions of aging of the desired object in the time interval from 0 to 10 years was obtained. The laws of probability distribution of face detection by artificial neural networks are selected.

Ivanov V.A., Smirnov A.A., Nikolaev D.A. The real probability of recognizing images of people's faces using artificial neural networks.  Radiotekhnika. 2022. V. 86. № 1. P. 55−60. DOI: https://doi.org/10.18127/j00338486-202201-09 (In Russian)

1. Ivanov V., Zvezhinskij S. Problemy fal'sifikacii foto- i videomaterialov na sovremennom jetape razvitija cifrovizacii. Tehnologii bezopasnosti. 2021. № 1 (In Russian).
2. Ivanov V., Zvezhinskij S. O razrabotke otechestvennogo special'nogo programmnogo obespechenija dlja vyjavlenija vmeshatel'stva v foto i video izobrazhenija. Tehnologii bezopasnosti. 2021. № 2 (In Russian).
3. Gmurman V.E. Teorija verojatnostej i matematicheskaja statistika: Ucheb. posobie dlja vuzov. M.: Vysshaja shkola. 2003. 479 s.  (In Russian).
4. Dvojris L., Ivanov V., Galev K. Imitacionnaja model' signalov i pomeh dlja testirovanija sistem obnaruzhenija. Radiotehnika. 2020  № 2. S. 14−19 (In Russian).
5. Dvojris L., Ivanov V., Krjukov I. Realizacija kriteriev obnaruzhenija i razlichenija signalov v srede MathCAD. Radiotehnika. 2021. № 2. S. 44−49 (In Russian).