Flow Image Statistics

From Endrov

In lack of a better name, these operations create local image statistics and features.

Contents

Average, Rectangle

Calculate local average.

Complexity O(#pixels)

Bilateral filter

Noise reducing Bilateral filter.

Complexity O(#pixels*pw*ph)

Entropy circle/rectangle

Local entropy.

Complexity O(#pixels r^2) or O(#pixels*pw*ph)

Kirsch filter

Edge detecting Kirsch operator.

Complexity O(#pixels)

Kuwahara filter

Noise reducing Kuwahara filter.

Complexity O(#pixels)

Percentile rectangle

Compute local percentile around every pixel. Percentile should be within [0,1] values outside will be clamped. Special cases:

  • 0 Min
  • 0.5 Median
  • 1 Max

Complexity O(#pixels*pw*ph)

Sum circle/rectangle

Local sum.

Complexity O(#pixels r^2) for circle and O(#pixels) for rectangle

Variance circle/rectangle

Local variance.

Complexity O(#pixels r^2) for circle and O(#pixels) for rectangle