A Face Feature Extraction and Classification Method

Abstract

The invention relates to a method for extracting and classifying human face features. The method comprises the following steps: using 2D-PCA method to perform feature dimensionality reduction on human face images, converting a high-dimensional image matrix into a low-dimensional image matrix; converting the low-dimensional The image matrix is ​​converted into a one-dimensional column vector; according to the one-dimensional column vector of the training set image, the intra-class scatter matrix SW and the inter-class scatter matrix SB of the training set are obtained, and the eigenvalue decomposition of SW and SB is performed respectively: use Dα Estimated by Dβ estimation can be obtained S ^ W - 1 = U W D α 2 U W T , S ^ B - 1 = U B D β 2 U B T ; Calculate the column space W1 and column space W2 respectively, and obtain the optimal projection space W=[W1,W2] of the feature extraction algorithm based on 2D-PCA two-level LDA; project the low-dimensional image matrix in (1) In the optimal projection space W, the feature vector of the image is obtained; the feature vector obtained in (6) is used for classifier training with the SVM+NDA model, and the final face classifier is obtained.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to a face feature extraction and classification method. Background technique [0002] The identification methods of the traditional time attendance system are mainly attendance card and radio frequency card. Due to the separability of the identification person, it is easy to cause the phenomenon of punching the card. Therefore, biometric identification technology has gradually become the main means of identification. At present, the fingerprint time attendance system using biometric identification technology has been widely used. However, the fingerprint attendance system needs special image acquisition equipment to acquire fingerprints, and the image acquisition is touch or contact, which will bring discomfort to users. Moreover, there are many groups or individuals whose fingerprint features are so few that it is difficult to image; when users use fingerprint collection equipment,...

AI Technical Summary

Problems solved by technology

However, the actual face recognition attendance system will also face some challenges. Affected by factors such as illumination, occlusion, scale or movement of the face area, the current face recognition has not yet achieved the expected effect

Images