A 4D expression recognition method based on multi-feature fusion based on tensor decomposition

Abstract

A 4D expression recognition method based on multi-feature fusion based on tensor decomposition, obtaining 4D facial expression data; after preprocessing the 4D facial expression data, the three components of the normal vector, the shape index and the depth of the 4D facial data are calculated. Figure; perform tensor decomposition on the three components of normal vector, shape index and depth map of 4D face data respectively, and extract dynamic facial expression information; use dynamic image network to classify dynamic facial expression information, and classify the classification The results are fused to obtain the final classification result. The invention makes full use of the information of the 4D face data, calculates the three components of the normal vector of the face, the shape index and the depth map for the sequence face data, and makes full use of the 3D geometric information of the face. It is said that the features of the face are more representative and discriminative, and the accuracy of face recognition and expression recognition is also higher.

Description

technical field [0001] The invention relates to an expression recognition method, in particular to a 4D expression recognition method based on tensor decomposition and multi-feature fusion. Background technique [0002] With the development and progress of artificial intelligence and computer technology, facial expression recognition and face recognition have attracted more and more attention. Expression recognition and face recognition are gradually becoming more and more widely used in life. There are many kinds of facial expression recognition methods at present, such as: using deep neural network to extract the features of 2D pictures or the features of videos, and then classify them. There are also expressions classification using 3D face data. In fact, 2D-based facial expression recognition is easily affected by lighting and scenes. 3D-based expression recognition can overcome the influence of lighting and posture, but for 3D expression recognition, because differen...

AI Technical Summary

Problems solved by technology

In fact, 2D-based facial expression recognition is easily affected by lighting and scenes

3D-based expression recognition can overcome the influence of lighting and posture, but for 3D expression recognition, because different people express expressions in different ways and degrees, even with the same expression, different individuals vary widely.

Images