Neural network expression recognition method based on graph structure

Abstract

The invention relates to the field of biological feature recognition, in particular to a neural network expression recognition method based on a graph structure, comprising the following steps: locating a plurality of key points of facial expression recognition; A filter is used to extract the texture feature vectors of each key point. Each extracted texture feature vector is taken as a node, eachnode is connected with each other to form a graph structure, the connecting lines between key points are taken as edges of the graph structure, geometric feature information between key points is expressed by Euclidean distance, and the graph structure is used for replacing an expression image; the graph structure is input into a BRNN neural network; the result of the expression classification isoutput at the last time step of the BRNN neural network using a plurality of iterations. The invention fully utilizes texture and geometric characteristic information of different expressions to solve the problem that the expression recognition accuracy is reduced due to interference information existing on the image.

Description

technical field [0001] The invention relates to the field of biometric feature recognition, in particular to a neural network expression recognition method based on a graph structure. Background technique [0002] In the research of facial expression recognition, extracting effective expression information from facial images is a key step. Early research on facial expression recognition mainly started from three aspects: feature learning, feature extraction and separator construction. First, people extract information about facial appearance or geometric shape changes from images or video series, representative methods such as LBP-TOP, HOG3D, DTAGN and STM-ExpLet. Then, select a subset of features that can effectively represent facial expressions, and finally, according to the extracted features, build an effective classifier to recognize facial expressions. But relatively speaking, the traditional classification learning method is difficult to achieve better results. [0...

AI Technical Summary

Problems solved by technology

[0005] The present invention intends to provide a neural network facial expression recognition method based on a graph structure, making full use of the texture and geometric feature information of different facial expressions, so as to solve the problem that the accuracy of facial expression recognition is reduced due to the interference information existing on the image

Images