Fractal Image Filters for Specialized Image Recognition Tasks
Abstract
Michael Barnsley made two visits to Wright Patterson to discuss opportunities for deployment of superfractals. In these lectures he argued that it is feasible to use superfractals in the design of antennas. This would have advantages over standard fractal designs: (i) implementation of V -variable models, whereby the structure of the antenna is of several characters at each level; (ii) simplicity of reconfiguration to different designs, for testing. The approach would be based on the use of fractal homeomorphisms. A number of papers related to this topic were written during the period of the grant. A high level programmer with Matlab expertise was retained to investigate the feasibility of making an image recognition system based on fractal homeomorphisms, in a Matlab environment. In view of the decision to focus on antenna design, the poor performance of Matlab code in comparison with Barnsley's existing code, and the expense of continuing to work with Matlab, this direction was discontinued. Some progress towards an intellectually satisfying model for living systems, using discrete superfractals, was made.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 11, 2010
- Accession Number
- ADA513998
Entities
People
- Andrew Vince
- David M Wilson
- Michael Barnsley
- Ross Atkins
Organizations
- Australian National University