Palm vein authentication method based on convolutional neural network

Abstract

The invention discloses a palm vein authentication method based on a convolutional neural network. The palm vein authentication method based on the convolutional neural network comprises the followingsteps of S1, training the convolutional neural network according to a training sample set image; S2, inputting a user registering image into a convolutional network model for generating a characteristic vector; S3, inputting a to-be-identified image; S4, performing comparison identification on a to-be-identified characteristic vector and a template characteristic vector which is stored in a template storage module; and S5, extracting a highest value of a probability value which is obtained from a result, if a highest value is larger than a fixed threshold, determining authentication success,and otherwise, determining authentication failure. According to the method, through the improved model, model scale is greatly reduced, and furthermore a plurality of data set combined training methods are supplied for settling a defect of small sample scale in identity authentication application. In a characteristic authentication step, multiple perception machines are utilized for calculating similarity probability of comparison identification. The parameters of the plurality of perception machines can be trained on line and updated automatically, thereby improving authentication speed and precision.

Description

technical field [0001] The invention relates to the technical field of vein recognition, in particular to a palm vein authentication method based on a convolutional neural network. Background technique [0002] With the development of technology, biometric identification technology is becoming more and more popular, and it is gradually replacing traditional identity authentication methods such as passwords, which brings great convenience to people. Traditional fingerprint recognition has been applied to various fields such as security, access control, and finance. Vein recognition technology has gradually attracted the attention of researchers and commercial companies due to its natural anti-counterfeiting properties. The palm veins are distributed under the surface of the skin, which belong to the internal physiological characteristics of the living palm, and have the characteristics of high security, uniqueness and strong anti-counterfeiting. [0003] Compared with finge...

AI Technical Summary

Problems solved by technology

At present, applications in identity authentication, such as fingerprints and veins, are applied in a small sample and a small range, and it is difficult to obtain a better recognition model through pre-training

Images