Crowd control with vector painting

Abstract

Producing crowd simulations is a time-consuming task. The amount of technical work required for controlling relevant crowd parameters is a hurdle for effective production. In this paper we propose a new authoring method that can easily create crowd simulations with intuitive painting interface. For this, Self Organized Distribution Map (SODM), Stroke-based Flow Field and Gradient Vector Flow (GVF) techniques were applied to automatically generate parameters from the image. A hybrid fluid model for crowd dynamics was then applied to create smooth crowd movement. Through this system, a user can quickly create the intended crowd simulation with various movements. This methodology would enable fast crowd simulation prototyping in an intuitive way. Copyright © 2014, Australian Computer Society Inc.