Face authentication method and device

Abstract

The invention provides a face authentication method and device, which belong to the field of image processing and pattern recognition. The method includes: using a Gabor filter to extract Gabor texture features from acquired face image samples, and separating the extracted Gabor texture features The real and imaginary parts are used to obtain several feature maps; the obtained feature maps are represented by LBP statistical histograms, and summarized as the face feature expression of the face image sample pair; the AdaBoost algorithm is used to express the face feature of the face image sample pair Perform feature selection; use PCA algorithm and LDA algorithm in turn to perform feature dimensionality reduction and feature discrimination processing on the face feature expression of the face image sample pair after feature selection. Compared with the prior art, the face authentication method of the present invention has the advantages of being able to fully provide sample texture information, requiring less sample quantity, and having low time and space complexity of the algorithm.

Description

technical field [0001] The invention relates to the fields of image processing and pattern recognition, in particular to a face authentication method and device. Background technique [0002] As the most effective document that can represent the identity of adults over 16 years old in China, the second-generation ID card records the holder's unique ID number, face image and other relevant information. However, the emergence of ID card theft, false ID cards, and multiple ID cards for one person has caused huge losses to many industries and countries. Therefore, using the second-generation ID card for identity authentication has become an important method to avoid the above phenomenon. This method is to compare the photo of the second-generation ID card with the face image collected on the spot to determine whether it is the same person. [0003] Due to the low resolution of the face image on the second-generation ID card, there is often a large age difference between the fac...

AI Technical Summary

Problems solved by technology

At present, ID card authentication methods in the prior art include methods based on eigenfaces and methods based on Gabor features, etc. However, for methods based on eigenfaces, the accuracy of face positioning is required to be high, and the method is not sensitive to occlusion, The change of posture is more sensitive, which can easily lead to inaccurate authentication; for the method based on Gabor features, more texture information of the image is required, and the number of samples required for training is larger, the corresponding time complexity is higher, and the dimension of the feature is obtained High, complex calculation, usually unable to meet real-time requirements

Images