From HIT Lab NZ
3D Facial Expression Analysis And Representations
The goal of this research is to investigate, develop and evaluate a generic mathematical framework for 3D facial expression modelling and animation. The motivation of this research comes from the need to control physics-based models in real-time using vision-based techniques. We plan to enhance vision-based animations with a new mathematical framework to control animation parameters. Virtual character animations based on this framework will be evaluated in a series of experiments that will compare user response to animated characters using our emotional based framework, and their response to more traditional characters.
Analysis and representation of a conceptual framework for 3D facial expression modelling and animation. Facial expression models are extensively used in real-time rendering and animation where interaction with virtual characters is present. Currently, are looking at the fundamental concepts in expression modelling and at the possible methodologies for extracting related facial features from motion captured data.
Physics-based modelling is computationally expensive and requires a vast number of control parameters in an animation. As a result this limits the use of physics-based models in real-time vision-based facial animations. This situation has motivated us to look at bridging the gab between vision-based animations and physics based modelling by proposing to establish a generic mathematical 3D facial expression framework which can be used to model and animate captured expressions.
- Development of generic mathematical models that can be used for 3D facial expression modelling and animations.
- A complete real-time system demonstrating the use of physics-based models in a vision-based facial animation using the developed methods.
- User evaluations of the system, which includes studying how emotions can enhance the 3D representation of a face model and its animation.
- A discussion of how the findings can enhance and support real-time vision-based facial animations using physics-based models.
Publications
Additional Content
Image Gallery
Videos
There are currently no videos uploaded...
