A Real Time System for Model-based Interpretation of the Dynamics of Facial Expressions

Recent progress in the field of Computer Vision allows intuitive interaction via speech, gesture or facial expressions between humans and technical systems.Model-based tech-niques facilitate accurately interpreting images with faces by exploiting a priori knowledge, such as shape and texture information. This renders them an inevitable component to realize the paradigm of intuitive human-machine interac-tion. Our demonstration shows model-based recognition of facial expressions in real-time via the state-of-the-art Candide-3 face model [1] as visible in Figure 1. This three-dimensional and deformable model is highly appropriate for real-world face interpretation applications. However, its complexity challenges the task of model fitting and we tackle this challenge with an algorithm that has been auto