Computability in Europe 2006
Logical Approaches to Computational Barriers


Regular Talk:
Two-Dimensional Cellular Automata Transforms for a Novel Edge Detection


Speaker: Seoktae Kim
Author(s): Yongri Piao, Seoktae Kim and Sung-Jin Cho
Slot: Array, 12:00-12:20, col. 5

Abstract

In this paper, we propose a novel edge detection scheme using two-dimensional
cellular automata transforms (CAT). Cellular automata (CA) are discrete
dynamical systems whose function is completely specified in terms of local
relation. First, we get the gateway values such as wolfram Rule, number of
cells in lattice, number of cells per neighborhood, initial configuration and
boundary configuration. Second, we use the gateway values to generate a
dual-state, two-dimensional and dual-coefficients basis function. Finally, we
transform images into cellular automata domain according to the basis function.
Then we use the basis function and cellular automata
transform coefficients to extract the edge of the image.  The experimental
results verify that the proposed method is a new attempt to detect edge with
cellular automata model as well as to maintain its high efficiency.

websites: Arnold Beckmann 2008-05-28