A Finger Vein Recognition Method Fused with Local Coding and CNN Model

Abstract

A Finger Vein Image Recognition Method Based on Fusion Local Coding and CNN Model. It includes extracting the ROI of the finger vein image; using the encoding operator based on the weighted symmetric local graph structure to encode the features of the image: obtaining the reconstructed encoding convolution filter; establishing an improved convolutional neural network model; Perform feature extraction on the ROI image; perform similarity measurement on the finger vein ROI image to be matched. The finger vein image recognition method fused with local coding and CNN model provided by the present invention can not only solve the problem of changing finger gestures to a certain extent, but also has a good matching effect, and reduces the number of parameters that can be learned by the network model, reducing the learning curve. The matching efficiency is improved, and the experimental results on two finger vein databases show that this method has certain feasibility.

Description

technical field [0001] The invention belongs to the technical field of finger vein image recognition, in particular to a finger vein image recognition method that integrates local coding and a CNN model. Background technique [0002] With the rapid development of computer technology and the advent of the information age, traditional biometric identification technology can no longer meet people's needs, and people's requirements for the accuracy of identification technology are getting higher and higher. Compared with other biometric identification technologies (such as: face, fingerprint, iris and palm print, etc.), finger vein has the advantages of vitality, uniqueness, user-friendliness and long-term invariance. In addition, since the finger vein is located under the skin, the image of the finger vein is usually collected using near-infrared (NIR) imaging, so the finger vein is non-contact and difficult to be copied and forged. In recent years, finger vein recognition tec...

AI Technical Summary

Problems solved by technology

[0003] Due to the large number of learnable parameters of the existing convolutional neural network model, the learning cost of the network model is high, and there may be a problem of overfitting

At present, people usually directly reduce the number of parameters of the neural network structure to reduce the learning cost, but this may cause the network performance to decrease accordingly.

Images