![]() | Only 14 pages are availabe for public view |
Abstract The aim of this thesis is to study a special type of modem analysis, Fractal Geometry, which has many applications. We emphasize on fractal application in image processing caUed fractal image compression. Image compression is needed in computers and modem technology for efficient storage and speed of transmission. We study a new feature offractal image compression, image enhancement. This feature is to be added to the known benefit of fractal image compression, the high compression ratio. To study fractal image compression, we introduce a fractal metric space. Using modem set theory of Georg Cantor, a complete metric space called Fractal Space or Hausdorff Space was established. The Contraction Mapping Theory assures the existence of fixed points for contractive affine transformation. Those fixed points were called the attractor of transformation. They show amazing images when drawn using computer program. But, it was not part of om work to study such images. We address the reverse problem, if we have an image, how can we find a transformation whose attractor is the given image. This is called fractal image encoding. This problem was first addressed and solved by Michael Bamsley. We use Yuval Fisher proposed quad-tree partitioning algorithm. We give a detailed discussion of quad-tree algorithm Emphasis is made on the advantages of applying fractal image compression method to enhance the image properties. We did not examine coding or decoding time, obtained compression ratio or complexity of technique since these areas are well covered in other searches |