Modeling emotion for anthromorphic agents

• Main Authors: Siraj, Fadzilah, Ab. Aziz, Azizi, Sainin, Mohd Shamrie
• Format: Conference or Workshop Item
• Language: English
• Published: 2005
• Subjects: QA76 Computer software
• Online Access: http://repo.uum.edu.my/1560/1/modeling_emotion_for_anthromorphic_agents_SEMINAR_KEB._SOSIOEKONOMI_DAN_IT_KE3%5B1%5D.pdf; http://repo.uum.edu.my/1560/

Emotions should play an important role in the design of interfaces since human beings interact with machines as if they were social actors. To investigate if and how machines can express emotions, it is necessary to investigate human-human interaction. Facial expressions are one of the most powerful, natural and immediate means by which human beings communicate their emotions and intentions. Face-to-face communication is inherently natural and social for human-human interactions. There is substantial evidence that suggests this may also be true for humancomputer interactions. In other words, human beings regard computers as social agents with whom “face-to-interface” interaction may be most easy and efficacious. Based on Ekman’s theory, SIX (6) universal emotion expressions that do not change too much from culture to culture were adopted in this study. The six emotion expressions are happiness, sadness, disgust, anger, surprise and fear. In addition, based on these six universal expressions, the AeMotion system was developed using Visual Basic 6.0. The Sony Digital Handycam video camcorder was used to capture the facial expressions. The system provides meaningful information for emotion detection through human facial expression. The final result of pixelization can be transferred into the set of processing array for emotion recognition purpose. The pixel formation provides indirect information for emotion (sad, anger, disgust, happy, fear and surprise) cues such as “brows lowered and drawn together” portrays action of disgust. The study demonstrates that the human facial expressions were successfully captured and pre-process to represent image-based emotions.The study also demonstrates that the image-based parameters could be used to interpret the facial affect space. In addition, the results also demonstrate that there is a need for both basic and applied research contributions to the rapid developing field of affective computing. Currently, experiments are still being conducted to see the impact of a variety of compressed image conditions has on affect space.