Presenters for 2004:   Stephanie Andrews   Christopher Bailey   Joanna Berzowska   Shushil Bhakar   and Eric Hortop   Cliff Burgess   Paulo Chagas   Dennis Dollens   Dan Falk   Sarah Filley   Ivette Fuentes-Guridi   Lila Kari   Narendra Pachkhede   Chris Salter   Chelsea Smock   Clara Ursitti   Derek van der Kooy   Yon Visell   Fabian Winkler   Panel Discussion Symposium, 2004 Generative Design Project By Concordia University researchers: Sushil Bhakar (CS doctoral candidate), Dr. Cheryl Kolak Dudek (Fine Arts), Eric Hortop (Math & Communications major), Dr. Sudhir Mudur (CS), Sylvain Muise (Math major), Dr. Lydia Sharman (Design) and Dr. Fred Szabo (Math). Presented by Sushil Bhakar and Eric Hortop Concordia University http://homepage.mac.com/generativedesign http://www.cs.concordia.ca/~sushi_bh/ http://www.metawidget.net Our research focuses on innovative applications of digital technology to the analysis of Kuba cloth and Zillij mosaics. Mathematical considerations led us to use neural networks, shape grammars, and related technologies. Our goal is to build an interactive database for the study of historical pattern and the generation of contemporary designs. Our aim is to develop a system for generating rules to describe groups of patterns. Zillij mosaics are symmetrical designs made up of repeating tile shapes. Color in the Zillij tiles reinforces a layered reading of pattern. Fractal symmetry is displayed in the Kuba designs in a non-symmetrical composition. Analysis of the rules defining various groups may allow for finer distinctions of classification, as well as assisting the design of new patterns. The jump from images of pattern to shape grammar rules requires tools to extract edges from raster images. We have approached the problem from both a hand tracing and automated perspective, pulling out groups of shapes via shape parameters and thresholding. The role of symmetry in discovering pattern and natural ways of providing stopping conditions for a shape grammar in our context have been concerns in developing automated rule extraction. Group theory leads us to a finite number of types of symmetry, easing the hunt for symmetry-based rules. Invisible marker elements have been the conventional solution in shape grammar implementations. A stopping condition based on the relative sizes of the smallest and largest features in a design, seems to be more realistic for computer extraction. Biographies: Sushil Bhakar is a PhD student in computer science at Concordia University in Montreal. His research interests are 3D computational graphics and visualization. His hobbies include abstract algebra, mathematical logic and their applications. His current projects are real-time point-based rendering of three-dimensional objects and the mathematical structure of patterns discussed in this presentation. Eric Hortop is a double-major co-op student in mathematics, statistics, and communication studies at Concordia University in Montreal. He plans to continue on in research and envisages a career in academia or government. His other interests include computers, graphic design, typography, cooking and long-distance bicyling.   Page Top Copyright © 2004 Subtle Technologies Festival