As humans, our fascination with recreating images of ourselves began with parietal art many thousands of years ago, and has continued throughout the ages. However, we also now strive to create moving, thinking, believable, virtual humans, rather than focus entirely on static images. Moreover, computer generated humans are now integral to the digital world that encompasses us, yet 3D animation remains rife with challenges. The overarching objective of the thesis, therefore, is to systematically investigate and design a comprehensive computational model that leverages the effect of natural motion from the real-world to mediate enhanced 3D animation production. Thus natural motion is simulated to support user engagement, leveraged through real-time mapping, and for control of digital characters through enabling devices for markerless capture. Captured data is analysed offline to identify potential creative enhancements. As a result of the comprehensive investigation, contributions include an innovative animation framework to support real and virtual engagement within immersive environments. The novel approach simulates motion, through combining elements of real-time game technology with facets of behaviour simulation from Embodied Conversational Agents. Contributions are also proposed towards bridging the gap between realistic motion capture and authoring more stylised 3D cartoon animations. Shortcomings of transferring natural motion to exaggerated cartoon styled animation are systematically studied and a solution based on parametric motion curve optimization is proposed. Intelligent reasoning for validation of temporal sequencing of animation assets using a structured knowledge model, further contributes towards animation production through facilitating sharing animation planning across different domains. The computational model, thus embodies the concept that through systematic investigation of synergies and differences between natural motion and keyframe animation, the benefits of both can be fused together to target more efficient, yet believable, and even creative animation authoring.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:765986 |
| Date | January 2017 |
| Creators | Rivera, Fiona MacDonald |
| Publisher | Queen Mary, University of London |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://qmro.qmul.ac.uk/xmlui/handle/123456789/25937 |
Page generated in 0.0017 seconds