Image coding using fractal parameters
Yusuke Suzuki, Hiroshi Sumiyoshi, Akira Miyauchi · 1992
A fractal based image coding method is discussed. The regions of an image are classified into two categories using fractal parameters. The regions, categorized 'not have to preserve the waveform of the image' are encoded using only fractal parameters. The other one categorized 'have to preserve the waveform of the image' are encoded using any former waveform coding method (DCT, VQ, etc.). Some of the regions categorized 'have to preserve the waveform of the image' contain region contour lines. These regions are classified into two categories using the complexity of the shape of contour lines. Then the regions with simple contours can be encoded using the same method for the other regions. Another one with complex contours can be encoded using both the waveform coding method and the fractal coding method. Barnsley's Random IFS (iterated function system) is applied to the regions that 'have to preserve the waveform of the image'.>