Sticky Pixels: Evolutionary Growth by Random Drop Ballistic Aggregation

Roberts, Jonathan (2001) Sticky Pixels: Evolutionary Growth by Random Drop Ballistic Aggregation. In: Eurographics UK 2001 Conference Proceedings. (Full text available)

PDF
Download (397kB)
[img]
Preview
Postscript
Download (9MB)
[img]
Preview

Abstract

Over the years many techniques have been developed for simulating and modelling trees, ferns, crystals and natural structures. Indeed, many complex and realistic images have been formed. Often, these rely on rule based systems to create the structure, they start with a simple form and progressively refine it into a more complex form by applying rules. We use the notion of Sticky Pixels to form textures. The pixels (or objects) move around the space, when they touch another object they stick together to form a larger cluster. The objects aggregate and stop at the place and position where they first touched. Such an aggregation generates neighbourhoods of pixels that form natural looking shapes. The pixels may randomly walk around (such as using Brownian motion), or be guided along pre-defined routes (often described as ballistic), to obtain different structures. We use a ballistic aggregation technique, where the particles are randomly dropped onto a canvas, migrate and stick onto the closest position of the nearest cluster. We present Sticky Pixels, explain different parameters and describe our algorithm.

Item Type: Conference or workshop item (Paper)
Uncontrolled keywords: Fractals, Evolutionary Art, Diffuse Limited Aggregation, Sticky Pixels
Subjects: Q Science > QA Mathematics (inc Computing science) > QA 76 Software, computer programming,
Divisions: Faculties > Science Technology and Medical Studies > School of Computing > Applied and Interdisciplinary Informatics Group
Depositing User: Mark Wheadon
Date Deposited: 24 Nov 2008 17:59
Last Modified: 06 Sep 2011 01:11
Resource URI: http://kar.kent.ac.uk/id/eprint/13627 (The current URI for this page, for reference purposes)
  • Depositors only (login required):

Downloads

Downloads per month over past year