pcontrast1quadtree

Performs quadtree (or octree) segmentation based on contrast criterion.

Synopsis

pcontrast1quadtree threshold [-m mask] [im_in|-] [rg_out|-]

Description

pcontrast1quadtree segments the input image im_in into homogenous regions. Homogeneous regions are regions that have an inner contrast ≤ threshold.

The principle of the algorithm is as follows:

At the begin consider the image as the first block.
If the block violates the uniformity predicate (i.e. inner contrast ≤ threshold) then split the block into four equally sized sub-blocks and then apply the algorithm recursively on each sub-blocks.

Therefore, the result is composed of rectangular regions.

The contrast is calculated from:

    contrast(R)= 1/N * SUM(max(C(s,t), t in V(s) and t in R))
and C(s,t)= | im_in[s] - im_in[t] |

where N is the number of pixels of the region R.

For 3D image, the output region map is composed of octree regions.

Parameters

threshold is the maximum contrast value to decide if a region is homogeneous or not. Values are from the gray scale of the input image im_in (eg., 0-255 for Uchar image).

Inputs

im_in: a grayscale image.

Outputs

rg_out: a region map.

Result

Returns the number of regions.

Examples

Builds the quadtree of tangram.pan:

   pcontrast1quadtree 10 tangram.pan a.pan

C++ prototype

Errc PContrast1Quadtree( const Img2duc &im_in, Reg2d &rg_out, Uchar threshold );

Version française

Segmentation d'une image par quadtree (ou octree) selon le contraste.

Author: Laurent Quesnel