The synthesis of Javanese woman's facial image on anger expression based on emotion regulation

The synthesis of human face has been used in many applications. In the human-machine dialogue systems, human face that looks similar to its original face is used as the visual representation from the computer agent which interacts with the human user. Making the image of face synthesis from the human face that looks natural is not the simple thing, because human facial expressions influenced by many things not always followed by the emotion they want to show. The Javanese women are the interesting case, because of their character as the person who upholds the unity and harmony values. Therefore, they always show respect and well-mannered in their speaking and attitude to other people, including their face expression, even if it is different with their feeling. This is because they do not want to make any conflicts. In this research, we represent on how the harmony value as the regulation value influences the Javanese woman facial expressions when they are angry. The harmony values itself can reduce the ones anger emotion until 26%. The ANFIS classifier is used to classify the anger emotion, with three stimulus appraisal values (angry, rather angry, and not angry) and one regulation value (harmony value) as the inputs. The output of this classifier is the synthesis of Javanese woman's face expression. Based on the research, we conclude that the Javanese woman face synthesis can be presented close to their real character. By the regulation value, the expression of the Javanese woman anger face will be reduced; therefore, its image will show not angry or hold their anger. This is compatible with the character of Javanese woman when they are angry, where they do not show over-reacting to avoid any conflicts.