Human face super-resolution reconstruction recognition method based on fractional order orthogonal partial least squares

Abstract

The invention discloses a human face super-resolution reconstruction recognition method based on fractional order orthogonal partial least squares. The method comprises the following steps: 1, readjusting intra-group and inter-group covariance matrixes through fractional order eigenvalues and singular values by using fractional order orthogonal partial least squares, calculating a projection direction, mapping face image features into subspaces, and reconstructing low-resolution input high-resolution global face features through a neighborhood reconstruction idea; 2, constructing a high-resolution face residual block by using a neighborhood reconstruction method, synthesizing the residual block to obtain high-resolution face residual compensation, and supplementing face details through a residual compensation strategy; and 3, adding residual compensation to the global face to form a high-resolution face image output by the final algorithm, wherein the high-resolution global face features can be used for face recognition. In the face super-resolution reconstruction and recognition application, a better face reconstruction effect and higher face recognition accuracy can be obtained.

Description

technical field [0001] The invention relates to the field of super-resolution reconstruction and recognition, in particular to a face super-resolution reconstruction and recognition method based on fractional order orthogonal partial least squares. Background technique [0002] Multivariate analysis methods are often used in super-resolution reconstruction for feature extraction, among which principal component analysis (Principal Component Analysis, PCA) is more popular, and the feature extraction step is usually used to reduce the dimensionality of data and reduce noise. PCA extracts the useful information of the face and filters the noise by retaining the appropriate dimensions. Wang et al. proposed a framework for generating high-resolution faces by obtaining linear combination coefficients of images through PCA. [0003] Partial least squares (PLS) is an effective method for analyzing the relationship between two types of random variables, which aims to find a pair of p...

AI Technical Summary

Problems solved by technology

Learning-based methods predict high-resolution images by learning the relationship between high-resolution and low-resolution training sets. Recently, many researchers have combined deep learning with learning-based super-resolution methods and achieved great success. ; Interpolation-based methods generate high-resolution images by predicting unknown pixel information, but because they do not bring effective new information, the results are usually very blurred; reconstruction-based methods use prior knowledge and constraint information to construct high-resolution images, but Still poor at detailing output

Images