org.openimaj.image.processing.convolution

Class FTriangleFilter

java.lang.Object

org.openimaj.image.processing.convolution.FTriangleFilter

All Implemented Interfaces:

ImageProcessor<FImage>, Processor<FImage>

public class FTriangleFilter
extends Object
implements ImageProcessor<FImage>

Image convolution with a triangular filter. Implementation is based on repeated convolution with rectangular kernels, which is done efficiently using a integral image style approach. Overall complexity is independent of filter size and linear in the number of pixels.

Author:

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

Constructor Summary

Constructors

Constructor and Description
FTriangleFilter(int filterWidth, int filterHeight) Construct with the given dimensions.
FTriangleFilter(int filterWidth, int filterHeight, boolean zeropad) Construct with the given dimensions.

Method Summary

Modifier and Type	Method and Description
static float[]	createKernel1D(int width) Construct a triangular kernel of the given size.
void	processImage(FImage image)

Methods inherited from class java.lang.Object

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

Constructor Detail

FTriangleFilter

public FTriangleFilter(int filterWidth,
                       int filterHeight,
                       boolean zeropad)

Construct with the given dimensions.

Parameters:

filterWidth - width of filter

filterHeight - height of filter

zeropad - zero-pad off the edge of the image if true; duplicate edge value otherwise.

FTriangleFilter

public FTriangleFilter(int filterWidth,
                       int filterHeight)

Construct with the given dimensions. Edge effects are handled by duplicating the edge pixels.

Parameters:

filterWidth - half-width of filter

filterHeight - half-height of filter

Method Detail

processImage

public void processImage(FImage image)

Specified by:

processImage in interface ImageProcessor<FImage>

createKernel1D

public static float[] createKernel1D(int width)

Construct a triangular kernel of the given size. The kernel will have 2*width - 1 elements.

Parameters:

width - the kernel half-width

Returns:

the triangular kernel