LBP human face light irradiation preprocess method based on Hamming distance restriction

Abstract

The invention relates to an LBP face illumination preprocessing method based on the limitation of hamming distance, which aims at effectively eliminating the influences of illumination variations in an image on face characteristics, improving the robustness of face recognition against illumination and achieving high recognition rate. The method adopts the technical proposal that: the method is realized by the support of a computer provided with matched digital image processing software; information relevant to a digital image is collected, sorted and modified; a pixel point threshold value is adjusted; encoding is carried out to each pixel point according to coordinates and gray-scale values; codes that are required to be adjusted are found out for repairing; the mode of each pixel point is predicted; the conversion of bidirectional modes is carried out selectively; the images are reverted, and results are output. The LBP face illumination preprocessing algorithm has the advantages of eliminating the influences of illumination variations on face recognition, effectively improving the robustness of face recognition against illumination and increasing recognition rate.

Description

technical field [0001] The invention belongs to the technical fields of image processing, computer vision and pattern recognition. In particular, it relates to an LBP face illumination preprocessing method based on Hamming distance constraints. Background technique [0002] In recent years, face recognition technology has received great attention and developed rapidly. Face recognition involves disciplines such as image processing, computer vision, pattern recognition, and artificial intelligence, and has broad application prospects in criminal investigation, public security, and other fields. Face recognition is a tricky and highly challenging research field. Changes in the intensity and angle of light and the defects of sensor imaging make the gray distribution of the face quite different, which reduces the recognition rate of the face. The recognition has a key influence, and the lighting preprocessing method to reduce and eliminate this influence is one of the main dif...

AI Technical Summary

Problems solved by technology

[0006] (2) Too sensitive to image noise: the comparison threshold in the LBP operator directly depends on the value of the center pixel, so it is very sensitive to the change of the gray value of the center pixel. When the image noise is serious, the calculated LBP value will produce a large error

Images