Sparse maintenance distance measurement-based human face identification method

Abstract

The present invention provides a sparse maintenance distance measurement-based human face identification method. The method comprises the steps of 1, extracting the human face data from all stored human face data and constructing a distance metric algorithm framework by utilizing the human face data based on the maximum boundary theory, wherein the extracted human face data are provided with the label information; 2, based on the sparse representation theory, mining the sparse structured information of samples and constructing a sparse weight matrix; 3, constructing a sparse preserving optimization function to maximally store the sparse structured information of samples in a newly constructed distance metric space; 4, by utilizing a regularization framework, integrating the maximum boundary theory with the sparse preserving optimization function to obtain a sparse preserving distance metric; 5, by utilizing a feature descriptor, extracting the image features of a to-be-identified human face and conducting the human face identification experiment in the sparse preserving distance metric so as to classify the data of tested faces. The method has the advantages of high identification accuracy and fewer parameters, which fully utilizes labeled samples and unlabeled samples.

Description

technical field [0001] The invention belongs to the technical field of data mining and artificial intelligence, and in particular relates to a face recognition method based on a sparse distance-keeping measure. Background technique [0002] Traditional identity verification methods—keys, passwords, etc., have the disadvantages of being cumbersome and easy to be stolen. With the development of the multimedia age, these traditional authentication methods have been gradually replaced. Face recognition technology is one of the most widely used authentication methods at present. To put it simply, face recognition technology is to process the face data to be classified, compare the correlation with the face data in the database, and recognize the captured face. In recent years, researchers have proposed many face recognition methods, based on face template matching, dimension reduction, neural network, SVM classifier and other methods have achieved good face recognition results ...

AI Technical Summary

Problems solved by technology

Many distance measurement algorithms (ITML, LMNN, etc.) in the prior art do not make full use of all unlabeled sample data, and most algorithms consider the category label of the sample and ignore the spatial position information between samples, making the distance measurement algorithm in many applications It is difficult to further improve the recognition accuracy

Images