class ipRelaxation

This class is designed to apply relaxation on the linear system induced by the discretization of an elliptic PDE (diffusion)

Inheritance:


Public Methods

[more] ipRelaxation(real lambda_, int type_, int steps_, int width_, int height_, const char* coding_)
constructor
[more]void process(Sequence* seq_in)
[more]void cutExtremum(real* data, int width, int height, int distribution_width)
This function truncates the pixel's distribution of an image: pixels whose value is below (or above) the specified distribution width are set to the new minimum (maximum) value (also defined by the specified distribution width).
[more]virtual ~ipRelaxation()


Inherited from ipCore:

Public Fields

output field

input/output image coding field

Public Methods

constructor and destructor

initialization and processing

accessing and modifying output


Documentation

This class is designed to apply relaxation on the linear system induced by the discretization of an elliptic PDE (diffusion)

Relaxation is an iterative method allowing the resolution (approximation) of large and sparse linear systems. Here, the Gauss-Seidel scheme with red-black ordering is used (see multigrid.h file). The number of relaxation steps can be provided (default 10).

o ipRelaxation(real lambda_, int type_, int steps_, int width_, int height_, const char* coding_)
constructor

Parameters:
lambda_ - relative importance of the smoothness constraint
type_ - type of diffusion (isotropic, anisotropic)
steps_ - number of relaxation steps to approximate the solution
width_ - is the width of the input image
height_ - is the height of the input image
coding_ - is the coding format of the input image

ovoid process(Sequence* seq_in)

ovoid cutExtremum(real* data, int width, int height, int distribution_width)
This function truncates the pixel's distribution of an image: pixels whose value is below (or above) the specified distribution width are set to the new minimum (maximum) value (also defined by the specified distribution width). Used for display purpose.

Parameters:
- data the data the truncate (image)
- width is the width of the input image
- height is the height of the input image
distribution_width - the new half-width of the pixels distribution: the new distribution is in the interval [mean - distribution_width*std_dev; mean + distribution_width*std_dev]

ovirtual ~ipRelaxation()


This class has no child classes.
Author:
Guillaume Heusch (heusch@idiap.ch)
Version:
2.0 Wed 28 Jun 2006 05:19:33 PM CEST
Since:
2.0

Alphabetic index HTML hierarchy of classes or Java



This page was generated with the help of DOC++.