Iris Image Classification Method Based on Deep Learning Features and Fisher Vector Coding Model

Abstract

The invention provides a method for classifying iris images, comprising: processing sample iris images in the construction stage of iris texture primitives to obtain iris texture primitives; A vector machine is used to construct an iris classifier; the iris classifier is used in the discrimination stage to classify the target iris image. The iris image classification method of the invention can effectively complete the iris image classification problem, and improves the efficiency and safety of iris recognition. The present invention uses the features obtained by deep learning to replace the traditional hand-designed features to extract iris texture primitives, which has the advantages of high precision, high robustness and high reliability, and is suitable for living body detection, race recognition, gender recognition, etc. Iris image classification problem for various application requirements. The invention effectively solves the problems of system safety, large-scale data retrieval and the like encountered in the process of commercialization of the iris system.

Description

technical field [0001] The invention relates to the technical fields of computer vision, pattern recognition and machine learning, in particular to a method for classifying iris images based on deep learning features and Fisher Vector coding models. Background technique [0002] With the rapid development of the Internet, the relationship between people is closer and the interaction is more frequent. Identification based on biometrics has attracted people's attention and has penetrated into every aspect of people's daily life. Among many biological characteristics, iris has the advantages of high uniqueness, strong stability, and non-invasiveness. These advantages make the iris particularly suitable for identification and identification of people. It has received more and more attention in the past ten years, and related research and technology have also developed rapidly. Iris recognition can not only be applied to e-commerce, financial securities, information security, tr...

AI Technical Summary

Problems solved by technology

[0006] With the continuous development of hardware and software, the iris recognition system is developing in the direction of user-friendliness and ease of use, and the requirements for user cooperation are reduced, which will lead to problems such as the degradation of iris image quality, which makes it difficult in the actual application system. Find the optimal most discriminative features

With the increase of classification categories, there is still room for improvement in existing iris classification methods, and how to quickly and efficiently classify in iris recognition systems is still a difficult problem

Images