Compact graph structure face feature extraction algorithm with multi-direction weight optimization

Abstract

The invention relates to a compact graph structure face feature extraction algorithm with multi-direction weight optimization, which comprises steps: in a face gray image, a neighborhood with the sizeof 5*5 is selected; graph structures in four directions are constructed in a center point, each graph structure generates characters of 0 or 1 in a difference form according to a certain sequence, and four binary strings are formed; the four strings are assigned with different weights respectively for weighing, and the binary strings are finally converted to decimal numbers; the maximum number inthe four decimal strings is taken as the feature value of a center pixel; and finally, three neighbor classifiers are used for classification. The design is reasonable; pixels around the center pointare combined thoroughly; the problem that feature extraction is not in place due to mutation of the image gray in a certain direction is avoided, the weight is reasonably distributed on the basis ofthe four obtained binary strings, the face image feature information can be described comprehensively, effective face recognition can be carried out, and the recognition rate is greatly improved.

Description

technical field [0001] The invention belongs to the technical field of face recognition, in particular to a multi-directional weight optimized compact graph structure face feature extraction algorithm (MOWCLGS). Background technique [0002] The complete face recognition technology mainly includes the following three steps: face image preprocessing, face image feature extraction and face image classification and recognition. Specifically, the face is first detected in a complex environment, and then the region of interest is extracted from the face image, and then the corresponding face image features are extracted, and finally the face image is matched and recognized, and the processing process is detailed. Such as Figure 13 shown. [0003] Feature extraction is a key part of the face recognition system. Its purpose is to find the most representative and unique effective feature information from the image. The quality of the extracted feature information will directly aff...

AI Technical Summary

Problems solved by technology

Algorithms based on global features are suitable for describing the entire image, but the problem is that the foreground and background cannot be distinguished, especially if the area of ​​interest is occluded, the global features will be destroyed

However, the above two methods still have problems such as insufficient feature extraction caused by sudden changes in image grayscale in a certain direction, and unbalanced feature extraction due to uneven weight distribution.

Images