Adaptive Smoothing

Home • Gallery • Tutorials • Download • Purchase • Site Map
 

Adaptive Smoothing Properties Page

The Fractal Science Kit fractal generator Adaptive Smoothing page is found under the Orbital / IFS / Strange Attractor page and holds properties controlling the adaptive smoothing applied to Orbital samples. Orbital fractals frequently result in an image that appears grainy or speckled. This is a natural result of the algorithm used to create them. You can reduce these artifacts by increasing the number of points in an orbit but this can lead to significant increases in the time required to produce the fractal. Two additional methods are available to improve the results: Anti-Aliasing and Adaptive Smoothing.

Adaptive smoothing tries to reduce speckling by blurring the image using a low-pass filter. The strength of the filter decreases as the density of the sample increases so blurring is minimized where the samples are well represented.

See also:

Adaptive Smoothing

Fractal Science Kit - Adaptive Smoothing

The Adaptive Smoothing section contains properties that control the adaptive smoothing algorithm. The Active property is used to enable/disable adaptive smoothing. The Type property is either Epanechnikov or Gaussian and controls the shape of the smoothing kernel. Epanechnikov is the stronger of the two.

Maximum Weight is one of the values:

  • 3x3 Kernel
  • 5x5 Kernel
  • 7x7 Kernel
  • 9x9 Kernel
  • 11x11 Kernel
  • 13x13 Kernel
  • 15x15 Kernel

Maximum Weight gives the size of the kernel at minimum density. The larger the Maximum Weight, the smoother the resulting image. The kernel size is reduced as the sample density increases so that at maximum density, smoothing is eliminated. Power provides control over the smoothing effect for a given weight. Values of Power between 0.5 and 1 reduce the smoothing effect and values between 1 and 2 increase the smoothing effect.

Orbit Point Count Normalization

Fractal Science Kit - Adaptive Smoothing Orbit Point Count Normalization

The Orbit Point Count Normalization section defines Data Normalization parameters to shape the sample density used in the calculation.

 

Copyright © 2004-2019 Ross Hilbert
All rights reserved