pzeboudj

Computes the goodness measure based on inter and intra-region contrast.

pzeboudj [-m mask] [rg_in|-] [im_in|-]

pzeboudj computes a goodness measure for quantitative evaluation of gray level image segmentation results as defined by R. Zeboudj^*.

The measure is based on the inter and intra-region contrast.
Contrast of the pixel s with its neighbor t in the image I is measured as follows:

   c(s,t) = | I(s)- I(t) | / L-1

with

       L is max(ims)-min(ims).

The inner contrast of region Ri is:

Ii=1/Ai * sum_Ri [ max{c(s,t), t in W(s) inter Ri} ]

The outer contrast of region Ri is:

Ei=1/li * sum Fi [ max{c(s,t), t in W(s), t not in Ri} ]

where Fi is the boundary of Ri and li the length of Fi.

The contrast of region Ri is:

C(Ri) = { 1 - Ii/Ei if 0 < Ii < Ei;
        { Ei        if Ii=0;
        { 0         otherwise;

Finally, the global contrast is:

Contrast = 1/A * sum [Ai.c(Ri)]

The result is a value in [0..1]. The higher the value of the Zeboudj measure is, the better the segmentation result should be.

Caution: Regions with label=0 are not considered for computing.

Returns a positive real value.
(Use pstatus to get this value).

Computes the zeboudj measure for a simple binarization segmentation process:

   pbinarization 80 1e30 tangram.pan i1.pan
   plabeling 8 i1.pan i2.pan
   pzeboudj i2.pan tangram.pan
   pstatus

Errc PZeboudj( const Reg2d &rg_in, const Img2duc &im_in );

Calcul du critère de qualité basé sur le contraste inter et intra-régions.

^*JP. Cocquerez, S. Philipp, "Analyse d'images: filtrage et segmentation", Masson, 1995.

Author: Régis Clouard