3D virtual models are ubiquitous, notably spread to the general public by entertainment applications such as animation cinema and video games, and more recently, through virtual and augmented reality, or personalized 3D design. Thus 3D modeling evolves towards the need for increasingly complex 3D content that must remain visually plausible, and further be interactively created and deformed through efficient approaches allowing high-level control. In this context, various 3D models fulfilling these requirements through the integration of a priori knowledge are introduced. These models use high level geometrical constraints, potentially interleaved with underlying lower level models based on simulation or procedural generation, leading to hybrid approaches. Two axes of contributions are presented. First, 3D shape creation, applied to developable surface design, hierarchical shape deformation, and medical applications. Secondly, 3D animation, applied to articulated characters, interactive crumpled paper model, and fluid animation sculpting.
HDR Univ. Grenoble.