FAVS: 3D Facial Animation According to Vietnamese Semantic Analysis

There have recently been many studies on face animation according to sound, but facial expressions have not yet accurately represented and clarified the semantic meaning of the text. Studies show that characters need to represent at least six basic emotions: happy, sad, fear, disgust, anger, surprise. However, creating for facial animation for virtual characters is time-consuming and requires high creativity. The main objective of this study is to create Facial Animations according to Vietnamese Semantics (FAVS) more easily. The method is based on the important numerical blendshapes of the 3D model. The input text after predicting the emotion will be passed to the lips synchronous and emotion animating to perform the 3D face animation. Do the comparison with 2 methods: create animation by direct control with real human face via webcam and using keyframe methods. Assess the emotional expression of 3D characters according to the above 3 approaches. Survey respondents were asked to recognize a 3D Virtually sensitive emotional pattern generated for each sentence of text input and give a confidence score for each sentence. Survey results show, negative emotions are the most recognizable, happy and excited are easily confused.