Human face recognition method and system based on sparse representation and mean hash

Abstract

The present invention discloses a human face recognition method and system based on sparse representation and mean hash. The method comprises the following steps: preprocessing a human face test sample and all human face training samples: converting a colored human face sample into a grayscale image, and normalizing the human face test sample and all the human face training samples; extracting a sparseness feature among the samples by using a sparse representation model, encoding the human face test sample into a linear combination of all the human face training samples, and calculating a sparse coefficient matrix of the human face test sample on all the human face training samples; extracting a spatial structural feature inside the human face sample, and calculating a mean hash feature of the human face test sample and each human face training sample; fusing the sparseness feature among face samples with the spatial structural feature inside the sample; and determining a final category of the human face test sample by using a reconstruction error between the mean hash feature of the human face test sample and a reconstructed human face test sample.

Description

technical field [0001] The invention relates to a face recognition method and system based on sparse representation and mean hash. Background technique [0002] Face recognition is one of the core contents in the field of computer vision. It is the process of feature extraction, classification and recognition of face images. Authentication. Face recognition is a complex synthesis that integrates pattern recognition, computer vision, image processing and other technologies. It can be used in security inspection, security monitoring, criminal investigation tracking, identity recognition and other fields, and has broad application prospects. However, due to the limitation of the number of face samples, intra-class differences, inter-class similarities, illumination changes and many other factors, how to effectively extract face features from small face sample data sets and efficiently realize face classification, Thereby improving the robustness of face recognition and the ac...

AI Technical Summary

Problems solved by technology

However, the sparse representation algorithm only considers the sparsity between samples and ignores the spatial structure information inside the sample. When the face image is affected by strong illumination changes, the traditional sparse representation will reduce the robustness of face recognition. The recognition accuracy will drop significantly

Images