Research on Virtual Face Mapping Method Based on Expression Classification

PubDate: December 2021

Teams: Haikou College of Economics

Writers: Wang Hairong

PDF: Research on Virtual Face Mapping Method Based on Expression Classification

Abstract

This paper proposes a virtual face mapping method based on facial expression classification. Accord to that face data collected by the camera, CNN is adopt to classify the collected faces, and the classification result is expressed in the form of probability, that is, the expression classification ratio obtain by CNN network. At the same time, 3D_max is used to reconstruct the 3D model of facial expression, and the offset matrix is weighted and summed to output the virtual face model according to the classification scale, thus constructing a face mapping model based on facial expression classification. Experimental results show that the virtual face has a strong sense of reality.