Unaligned facial feature extraction method, device, computer device and storage medium

Abstract

The invention relates to the field of face recognition, in particular to a misaligned face feature extraction method, a device, a computer device and a storage medium. The method includes: obtaining aface image to be recognized; detecting symmetry of the face image; partitioning asymmetric face images into regions; carrying out feature extraction for each divided area. The invention extracts eigenvalues after segmentation of unaligned face images, when a human face moves horizontally or vertically in a picture. It can keep the regional feature unchanged or reduce the influence of human face misalignment on feature extraction, indirectly play the role of human face alignment, face feature extraction is more accurate, PCA dimension reduction processing is used to reduce the dimension of feature vector, reduce the computational difficulty, and improve the speed of face feature extraction.

Description

technical field [0001] The present invention relates to the field of face recognition, in particular to a non-aligned face feature extraction method, device, computer equipment and storage medium. Background technique [0002] With the rapid development and progress of computer science and technology, the application of face recognition technology is becoming more and more extensive. As an important part of biometric recognition technology, face recognition technology has high security, reliability and effectiveness. . [0003] Existing face recognition technology extracts the features of the face, and then recognizes the extracted features. The existing feature extraction methods include global feature extraction and regional feature extraction, all of which are based on the grayscale of the face All the pixel information in the picture is stored and then down, and then the LBP feature value of the image pixel is calculated, and the histogram statistics are performed accor...

AI Technical Summary

Problems solved by technology

[0003] Existing face recognition technology extracts the features of the face, and then recognizes the extracted features. The existing feature extraction methods include global feature extraction and regional feature extraction, all of which are based on the grayscale of the face All the pixel information in the picture is stored and then down, and then the LBP feature value of the image pixel is calculated, and the histogram statistics are performed according to these LBP feature values. However, all the pixel points are stored and calculated, and the calculation amount is too large, which wastes hardware resources. , the amount of data is too large, and the resulting histogram is too sparse, which affects the recognition speed, and the face feature extraction in the case of non-alignment is not accurate, and the face recognition is not accurate at that time

Images