Facial expression recognition method based on feature point vectors and texture deformation energy parameter

Abstract

The invention provides a facial expression recognition method based on feature point vectors and a texture deformation energy parameter. The method comprises the following steps: 1, respectively carrying out feature point positioning on neutral expressions and extreme expressions at a beginning end of a facial expression sequence by an AAM (Automatic Acoustic Management) tool of an OPENCV (Open Source Computer Vision Library); 2, forming feature point vectors with the selected 26 feature points to suit facial expression recognition, dividing the feature point vectors into a Euclidean distance d (representing a size) between the feature points and an included angle alpha (representing a direction) of connecting lines, calculating a distance coefficient ratio kd of the feature points according to d and alpha, and subtracting a redundant part kl to obtain kd - final (k alpha-final can be obtained in the same way); 3, establishing a feature block according to the feature points, calculating a texture deformation energy coefficient matrix, and finally obtaining the texture deformation energy parameter ks - final by PCA (Principal Component Analysis); and 4, inputting final features, namely using k final = kd - final + k alpha - final + ks - final as the training data of an RBF (Radial Basis Function) nerve network, and finally realizing facial expression recognition.

Description

(1) Technical field [0001] The invention relates to a facial expression recognition method, in particular to a facial expression feature extraction method based on feature points and feature blocks, belonging to the field of facial expression recognition. (2) Background technology [0002] Emotion recognition has always been an important research direction in the field of human-computer interaction. In order to establish a friendly and harmonious human-computer interaction mode, a large number of researchers start with voice, facial expression, text, etc., expecting to achieve better human-computer interaction effects with a single-mode or multi-mode fusion method. Among them, facial expression recognition is an important research direction of emotion recognition. In recent years, a large number of excellent research results have emerged, which undoubtedly strongly promoted the rapid development of human-computer interaction research. Thanks to the unremitting efforts of ex...

AI Technical Summary

Problems solved by technology

[0003] At present, in the field of facial expression recognition, combining facial structural features and texture information, and considering the impact of individual differences on facial expression recognition results, such methods are still relatively small, which restricts the research of facial expression recognition to a certain extent. Further development

Images