Parallel array grammars as models for the growth of planar patterns

We review the concept of growth on a discrete planar grid based on eight-ne ighbor time delays, and show that this process can be regarded as pattern g eneration by a context-sensitive isometric array grammar. We introduce a co ncept of growth in an environment by allowing the generation process to be influenced by a background pattern of scalar values, and illustrate how thi s increases the power of the model. We also demonstrate: in the domain of M R images of MS lesions, how the environment in which a growth process occur s can be recovered from an image. (C) 1999 Pattern Recognition Society. Pub lished by Elsevier Science Ltd. All rights reserved.