Compound Facial Expressions Generation with Arithmetical Features Space

Win Shwe Sin Khine

Japan Advanced Institute of Science and Technology (JAIST), Japan

Facial Expressions are the most common visual signal we have seen in our daily lives. Sometimes they convey emotions. For example, when human beings feel happy, they smile, and the facial expression related to smiles appear on their faces. Based on those expressions, we can realize human emotions. Research for facial expressions analysis has been developed in examining human behaviors based on visual signals; however, their focus is on simple human emotions, including happiness sadness. Simple emotions are a part of human emotions and cannot represent the whole human emotions. There is also a possibility that compound emotions have existed as a part of human emotions in a specific scenario. For instance, when we ride the roller-coaster, we feel fear and happiness simultaneously while riding. Like that circumstance, there are possible compound emotions that are required to be explored to get a deeper understanding of human complex emotions. Therefore, we will discuss compound emotions in this work with an analysis-by-synthesis approach considering the expressions features space.

Win Shwe Sin Khine

Japan Advanced Institute of Science and Technology (JAIST), Japan

Research Interest: Human Emotions; Facial Expression Analysis; Deep Learning;

WIN Shwe Sin Khine is a Ph.D. candidate at the Computer Imaging Laboratory from the Japan Advanced Institute of Science and Technology (JAIST). She received an M.Sc. in Information Science from JAIST in 2020 and B.Sc. in Computer Science from the University of Information Technology (UIT), Myanmar, in 2017. She is currently working on the research project for human complex emotions from facial expressions signals by analysis-by-synthesis approach to understanding the human deep emotions.

Organization