Finger vein recognition method based on convolutional neural network and supervised discrete hash algorithm

Abstract

The invention discloses a finger vein recognition method based on a convolutional neural network and supervised discrete hashing. The finger vein recognition method includes the steps: collecting finger vein images through bilateral infrared irradiation, and carrying out preprocessing, edge detection, direction correction and ROI extraction; extracting features of the ROI finger vein image by adopting a Res2net convolutional neural network; carrying out binary coding by adopting a discrete hash algorithm model; taking the extracted binary coding features as finger vein features to be registered / identified; and constructing a finger vein image database, carrying out the above processing of the to-be-recognized finger vein image, obtaining the recognized feature codes, carrying out the retrieval and recognition of the feature codes in the finger vein database one by one, obtaining the similarity through the Hamming distance measurement, and obtaining a matching recognition result. In theembodiment of the invention, feature codes with higher representation capability can be obtained through deep learning, and the discrete hash algorithm can enable the feature template to be smaller in size, so that the finger vein recognition method is more efficient in massive matching face matching.

Description

technical field [0001] The present invention relates to the fields of biological feature recognition technology, image recognition, pattern recognition and deep learning, in particular to a finger vein recognition method based on a convolutional neural network and a supervised discrete hash algorithm. Background technique [0002] With the rapid development of information technology, the use of biometric technology for personal identity verification has attracted more and more attention. Compared with traditional identity authentication methods, biometric identification technology utilizes the physiological or behavioral characteristics of the human body for identity authentication, which has extremely high security and will not be lost. The biological characteristics of the human body mainly include the face, iris, fingerprints, palm prints, finger veins, palm veins, etc., and the behavioral characteristics mainly include gait, handwriting, etc. Among these biometric featu...

AI Technical Summary

Problems solved by technology

[0004] At present, the problems that need to be solved are: the traditional image algorithm is not robust enough to extract manual features, which makes it difficult to improve the accuracy of recognition; the query speed in the massive finger vein database is slow and difficult to real-time, and the feature storage is limited in practical applications. The cost is also relatively high and the matching speed is slow

Images