Human facial expression mining based on cloud model

Abstract

In order to make the human facial expression recognition and analysis effectively, cloud model is used to mine the knowledge of human facial expression in this paper. Backward cloud generator without certainty degree treats the original input image as a cloud droplet image. The outputs are used to describe the three numerical characteristics-(Ex, En, He). How the three numerical characteristics of the cloud model (Ex, En, He) to embody the knowledge of facial expression? With two different sets of original images, two experiments are made to explain the specific meaning of the three numerical characteristics (Ex, En, He). Under the experiments, the profound knowledge may be extracted on the facial expression, i.e. Ex is the common of standard face, En is the difference, and He is the range of the difference.