de.unihalle.informatik.MiToBo.morphology

Class BinaryImageEndpointTools

java.lang.Object

de.unihalle.informatik.MiToBo.morphology.BinaryImageEndpointTools

public class BinaryImageEndpointTools
extends Object

Tool functions for extracting endpoints and branches from binary images, particularly images of skeletons.

Author:

moeller

Constructor Summary

Constructors

Constructor and Description
BinaryImageEndpointTools()

Method Summary

Modifier and Type	Method and Description
static Vector<Vector<Point2D.Double>>	findEndpointBranches(MTBImageByte img) Extracts set of endpoint branches from a given binary image.
static Vector<Point2D.Double>	findEndpoints(MTBImageByte img) Extracts set of endpoints from a given binary image.
private static void	traceBranch(MTBImageByte img, Vector<Point2D.Double> branch, Point2D.Double p) Continues a given branch at the given pixel recursively.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

BinaryImageEndpointTools

public BinaryImageEndpointTools()

Method Detail

findEndpoints

public static Vector<Point2D.Double> findEndpoints(MTBImageByte img)

Extracts set of endpoints from a given binary image.

An endpoint (e.g., of a skeleton) is defined as a foreground pixel having exactly one neighbor or having two neighbors which are located side by side. For determining the endpoints in an image the 8-neighborhood of all foreground pixels is checked.

Parameters:

img - Image to analyze (background black, skeleton white).

Returns:

Set of endpoints.

findEndpointBranches

public static Vector<Vector<Point2D.Double>> findEndpointBranches(MTBImageByte img)

Extracts set of endpoint branches from a given binary image.

An endpoint branch (e.g., of a skeleton) is defined as a sequence of pixels starting at an endpoint (see findEndpoints(MTBImageByte) for details) and ending up at the next branch point. In this context a branch point is a pixel with more than two neighbors.

Note that strict 8-neighborhood is assumed within the branch paths. Particularly, a branch point is assumed if a pixel has more than two neighbors. If the 8-neighborhood assumption is broken, a pixel not being a branch point can also have three neighbors. In such a case the algorithm will fail!

Parameters:

img - Image to analyze (background black, foreground white).

Returns:

Endpoint branches, first point of each list is endpoint.

traceBranch

private static void traceBranch(MTBImageByte img,
                                Vector<Point2D.Double> branch,
                                Point2D.Double p)

Continues a given branch at the given pixel recursively.

Note that strict 8-neighborhood is assumed in the path. If this assumption is broken and a pixel in the middle of the path which is not a branch point has more than two neighbors, the tracing will fail!

Parameters:

img - Image to analyze.

branch - List of already detected points of the branch.

p - Next branch pixel to examine.