Content-aware exaggerated editing for life-like captured animations

  • Authors:
  • Yann Savoye

  • Affiliations:
  • -

  • Venue:
  • Proceedings of the 9th European Conference on Visual Media Production
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Cutting-edge efforts have been invested in the automatic production of breath-taking visual effects involving time-varying data captured from real-actor performances. However, a challenges for computer-generated imagery is the puppetry of heterogeneous captured data, without the heavy use of artistic skills. Then, we focus on achieving desired exaggerated animations coherently while preserving life-life baked-in visual cues. In this paper, we propose a new method to generate content-aware exaggerated animations by melting motion, shape and appearance properties from captured data. In particular, our suggested approach explores two closely tools that serve the common theme of Animation-Cartoonization. The first one consists in realizing articulated-based stretchable cartoon editing from marker-based mocap clips. The second one generates video-based toon character from surface performance capture. Finally, we demonstrate the flexibility and stability of our approach on a variety of captured animations as input.