Finger vein recognition method and system based on cosine center loss

Abstract

The invention discloses a finger vein recognition method and system based on cosine center loss. The system comprises an image acquisition module, an image preprocessing module, an image recognition module and an image training module. The identification method comprises the following steps: acquiring a finger vein image of a user to be identified; performing image preprocessing on finger vein image information, and extracting a finger vein region of interest (ROI) image; and extracting finger vein features in the finger vein region of interest through a convolutional neural network based on cosine center loss, and identifying identity information of a to-be-identified user. The finger vein features can be effectively extracted, the redundancy of noise is improved, and the recognition precision of a finger vein recognition system is obviously improved.

Description

technical field [0001] The invention relates to a biometric identification method, in particular to a method and system for identifying finger veins based on cosine center loss. Background technique [0002] In the medical research of observing blood flow, people found that the blood flowing in the finger can absorb light of a specific wavelength, and a clear image of the finger vein can be obtained by irradiating the finger with light of a specific wavelength. [0003] Further medical research has proved that the shape of finger veins is unique and stable, that is, the image of each person's finger veins is different, and the images of different finger veins of the same person are also different, and the structure and distribution of finger veins will not change after adulthood. change. The above medical scientific research discoveries have laid a scientific theoretical foundation for the use of finger vein technology for biometric identification. [0004] Based on the in...

AI Technical Summary

Problems solved by technology

[0008] 1) It is difficult to obtain finger vein images, and it is difficult to establish an effective mathematical model and extract appropriate finger vein features under the current database;

[0009] 2) The vein image is susceptible to noise interference which affects the distribution related to the vein features in the image;

[0010] 3) The scarcity of finger vein samples leads to the lack of specificity of the current hand-designed finger vein features

[0011] Therefore, the current method based on manual features is difficult to effectively extract finger vein pattern information, resulting in limited recognition performance of the authentication system.

[0012] At the same time, the current application of deep learning in finger vein recognition also has great limitations, and it can only be applied to the recognition of large sample finger vein data and used as an image processing tool.

Traditional deep learning methods based on the softmax function are often limited to improving the inter-class distance of samples and classifying samples based on this method. However, this approach ignores the intra-class distance of samples, making it often requires large samples of finger vein data for training. , and did not give full play to its powerful feature learning ability

Images