org.openimaj.image.segmentation

Class KMSpatialColourSegmenter

java.lang.Object

org.openimaj.image.segmentation.KMColourSegmenter

org.openimaj.image.segmentation.KMSpatialColourSegmenter

All Implemented Interfaces:

Segmenter<MBFImage>

public class KMSpatialColourSegmenter
extends KMColourSegmenter

Simple image segmentation from grouping colours with k-means, and also incorporating a spatial aspect based on pixel location.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Field Summary

Fields inherited from class org.openimaj.image.segmentation.KMColourSegmenter

colourSpace, kmeans, scaling

Constructor Summary

Constructors

Constructor and Description
KMSpatialColourSegmenter(ColourSpace colourSpace, float[] scaling, int K, FloatFVComparator distance) Construct using the given colour space, number of segments, and distance measure.
KMSpatialColourSegmenter(ColourSpace colourSpace, float[] scaling, int K, FloatFVComparator distance, int maxIters) Construct using the given colour space, number of segments, and distance measure.
KMSpatialColourSegmenter(ColourSpace colourSpace, int K) Construct using the given colour space and number of segments.
KMSpatialColourSegmenter(ColourSpace colourSpace, int K, FloatFVComparator distance) Construct using the given colour space, number of segments, and distance measure.

Method Summary

Modifier and Type	Method and Description
protected float[][]	imageToVector(MBFImage image)

Methods inherited from class org.openimaj.image.segmentation.KMColourSegmenter

segment

Methods inherited from class java.lang.Object

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

Constructor Detail

KMSpatialColourSegmenter

public KMSpatialColourSegmenter(ColourSpace colourSpace,
                                int K)

Construct using the given colour space and number of segments. Euclidean distance is used, and the elements of each colour band are unscaled. Up to 100 K-Means iterations will be performed.

Parameters:

colourSpace - the colour space

K - the number of segments

KMSpatialColourSegmenter

public KMSpatialColourSegmenter(ColourSpace colourSpace,
                                int K,
                                FloatFVComparator distance)

Construct using the given colour space, number of segments, and distance measure. The elements of each colour band are unscaled. Up to 100 K-Means iterations will be performed.

Parameters:

colourSpace - the colour space

K - the number of segments

distance - the distance measure

KMSpatialColourSegmenter

public KMSpatialColourSegmenter(ColourSpace colourSpace,
                                float[] scaling,
                                int K,
                                FloatFVComparator distance)

Construct using the given colour space, number of segments, and distance measure. The elements of each colour band are by the corresponding elements in the given scaling vector; the scaling vector should be two elements longer than the number of colour bands of the target colour space such that the last two elements correspond to the scalings for the normalised x and y positions of the pixels. Up to 100 K-Means iterations will be performed.

Parameters:

colourSpace - the colour space

scaling - the scaling vector

K - the number of segments

distance - the distance measure

KMSpatialColourSegmenter

public KMSpatialColourSegmenter(ColourSpace colourSpace,
                                float[] scaling,
                                int K,
                                FloatFVComparator distance,
                                int maxIters)

Construct using the given colour space, number of segments, and distance measure. The elements of each colour band are by the corresponding elements in the given scaling vector; the scaling vector should be two elements longer than the number of colour bands of the target colour space such that the last two elements correspond to the scalings for the normalised x and y positions of the pixels. The k-means algorithm will iterate at most maxIters times.

Parameters:

colourSpace - the colour space

scaling - the scaling vector

K - the number of segments

distance - the distance measure

maxIters - the maximum number of iterations to perform

Method Detail

imageToVector

protected float[][] imageToVector(MBFImage image)

Overrides:

imageToVector in class KMColourSegmenter