Controllable texture synthesis for runtime simulation of large-scale vegetation

  • Authors:
  • Jing Fan;Jian-wei Ren;Ying Tang

  • Affiliations:
  • Zhejiang University of Technology, Hangzhou, China;Zhejiang University of Technology, Hangzhou, China;Zhejiang University of Technology, Hangzhou, China

  • Venue:
  • Proceedings of the 8th International Conference on Virtual Reality Continuum and its Applications in Industry
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The simulation of large-scale vegetation has always been important to computer graphics, as well as to virtual plantation in ecology. Although we can employ texture synthesis technique to produce large-scale vegetation from a small input sample, the synthesized result lacks physical accuracy which only focuses on visual appearance. On the other hand, the model developed in bio-physical area to simulate vegetation evolution provides meaningful distribution result of vegetation. However, the visualization of these meaning results is simple and preliminary. In this paper, we propose a new controllable texture synthesis method to generate visually-pleasing large-scale vegetation of physical meaning under the guidance of the control map. The control map is computed by the vegetation evolution model as well as considering the convenience to guide synthesis. Furthermore, we adapt our method on GPU to realize real-time simulation of large-scale vegetation. The experimental results demonstrate that our method is effective and efficient.