| Popis: |
Facial expressions play a significant role in conveying emotions with a widespread use across diverse cultures and societies globally. In particular, the expressions anger, sadness, fear, disgust, surprise, happiness and also neutral are considered universal. 2D and 3D avatar models are used to simulate facial expressions and have different applications in many domains. In this work, we consider a 3D model with facial expressions as a platform to analyze the basic set of expressions. We considered direction weighted intensity values of the FACS Action Units (i.e., also referred here as shape keys) relative to the nose tip, serving as a reference point, to generate direction weighted score for each target expression. The scores also give numerical validations for the repeated correlations indicated between a specific set of expressions (i.e., anger vs. sadness, and fear vs. disgust) in other research works that focus on developing techniques for facial expressions recognition and classification. In addition, the normal distribution of these seven expressions was depicted and gave a close to bell-curve shape which is an indication of a common phenomenon in nature. |
