FIELD: data processing.
SUBSTANCE: invention relates to biometrics, namely, to the technology for protecting various objects from access by unauthorised persons by means of personal identification based on the hand vein pattern (HVP). A method for personal identification based on the hand vein pattern, including stream reading of the hand vein pattern of the subject of identification, retrieval of a frame from the stream, analysis of the obtained frame, the fact of biometric identification, wherein, prior to the stream reading of the hand vein pattern of the subject of identification, the target subject of identification and the corresponding neural network are selected, and after receiving the frame from the stream, pre-processing thereof is conducted without changing the type of data, wherein the obtained frame is analysed by a convolutional neural network pre-trained for binary classification of images of a known format into two classes, wherein one of said classes indicates the correspondence between the frame and the selected subject of identification and the second class indicates the non-correspondence between the frame and the selected subject identification, whereas the result of the analysis is the vector of probabilities of the attribution of the selected image to each of the two named classes: Pi=(pti, p,fi), |I|=n, wherein I is the set of subjects i wherein the probabilities of neural network models predicting the probabilities of attribution of the frame to the subject of identification Pi correspond to said subjects; n is the power of the sets I, pti is the probability of the frame being attributed to the subject of identification i, Rfi is the probability of the frame not being attributed to the subject of identification i; wherein the fact of biometric identification is set in accordance with the condition pti ≥ Papprove.
EFFECT: increase the accuracy and speed of personal identification.
6 cl
Title | Year | Author | Number |
---|---|---|---|
METHOD FOR IDENTIFICATION OF PERSON BY PATTERN OF VEINS OF PALM | 2020 |
|
RU2788590C1 |
METHOD FOR TRACKING, DETECTION AND IDENTIFICATION OF OBJECTS OF INTEREST AND AUTONOMOUS DEVICE WITH PROTECTION FROM COPYING AND HACKING FOR THEIR IMPLEMENTATION | 2021 |
|
RU2789609C1 |
OBJECT RECOGNITION METHOD IN VIDEO SURVEILLANCE SYSTEM | 2022 |
|
RU2788301C1 |
METHOD FOR PERSONALITY AUTHENTICATION BY FINGERPRINTS AND DEVICE FOR ITS IMPLEMENTATION | 2007 |
|
RU2355307C2 |
METHOD OF PERSON IDENTIFICATION BY EEG-RESPONSE TO AMBIGUOUS IMAGES | 2017 |
|
RU2653239C1 |
METHOD OF CLASSIFYING FINGERPRINT PICTURES DURING PERSONAL AUTHENTICATION | 2011 |
|
RU2473125C1 |
AUTOMATIC PERSON IDENTIFICATION METHOD | 1998 |
|
RU2161826C2 |
METHOD FOR AUDIOVISUAL RECOGNITION OF PERSONAL PROTECTION EQUIPMENT ON HUMAN FACE | 2022 |
|
RU2791415C1 |
INTELLIGENT SYSTEM FOR DETECTING AND PREDICTING EVENTS BASED ON NEURAL NETWORKS | 2021 |
|
RU2797748C2 |
METHOD FOR HUMAN IDENTIFICATION BY HIS BIOMETRICAL IMAGE | 2005 |
|
RU2292079C2 |
Authors
Dates
2021-12-13—Published
2020-10-14—Filed