Matrix-based Joint Sparse Locality Preserving Projection Face Recognition Method

Abstract

The present invention provides a face recognition method based on a matrix-based joint sparse locality-preserving projection. Since the kernel norm measures the sum of singular values ​​of the matrix, it has stronger robustness to data changes, so the present invention utilizes the kernel norm The number is used as a basic metric to enhance the robustness of 2D‑LPP. Traditional feature extraction methods do not consider the sparsity of the projection matrix. Although based on L 1 The method can learn sparse projections, but it does not have the property of joint sparsity. This method utilizes L with the joint sparsity property 2,1 As a regular term, the norm can effectively learn the joint sparse discriminant projection, so as to perform effective feature extraction and selection. Experiments prove that the method of the present invention can effectively improve the robustness existing in 2D‑LPP and realize joint sparse feature extraction and selection.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a matrix-based joint sparse locality-preserving projection face recognition method. Background technique [0002] Research on face recognition began in the 1960s. At that time, due to limited computer processing power and immature algorithms, the capabilities of face recognition systems were very limited. After the 1980s, with the development of computer and imaging technology, face recognition technology has been continuously developed. In the context of the technological explosion, more and more people have studied this topic in the past ten years. Face recognition has shown great promise in human-computer interaction and public place safety monitoring. In automatic processing, digital image processing is an important part, and in terms of artificial intelligence, many functions are realized through digital image processing. [0003] Feature extraction is an importa...

AI Technical Summary

Problems solved by technology

[0005] The traditional learning model introduced above is not suitable for some feature extraction problems. The reasons are as follows: ①These three algorithms do not consider the influence of other factors on face recognition, such as shadows, lighting, etc., so its recognition under these factors The rate will decrease, so we use the kernel norm to complete the work of measuring the similarity between images

Images