VizKit::VisualConvolutionFilter Class Reference

Convolution filter for bitmap image effects. More...

#include <VisualConvolutionFilter.h>

List of all members.

Public Types
enum	Effect {   kNone = 0, kBlur, kMotionBlur, kGaussianBlur,   kLaplacian, kEdgeDetect, kSharpen, kEmboss }
	Constants for various convolution filters. More...
enum	EdgeAction { kZeroEdges = 0, kClampEdges, kWrapEdges }
	How to calculate edge values. More...
Public Member Functions
	VisualConvolutionFilter (const Effect effect, const uint8 colCount=3, const uint8 rowCount=3)
	The constructor.
	~VisualConvolutionFilter ()
	The destructor.
	VisualConvolutionFilter (const VisualConvolutionFilter &other)
	Copy constructor.
VisualConvolutionFilter &	operator= (const VisualConvolutionFilter &other)
	Assignment operator.
uint8	getNumberOfKernelValueColumns (void) const
	Returns the number of kernel value columns.
uint8	getNumberOfKernelValueRows (void) const
	Returns the number of kernel value rows.
float	getPostConvolutionScaleFactor (void) const
	Returns the post convolution scale factor.
float	getPostConvolutionBias (void) const
	Returns the post convolution bias.
void	copyKernelValues (float *buffer) const
	Copies the kernel values of the convolution filter into the provided buffer.
void	applyToPixelData (const PixelColor *pixelData, uint32 imageWidth, uint32 imageHeight, int pixelFormat, int pixelDataType, PixelColor **outPixelData, uint16 numberOfInterations=1) const
	Applies the convolution filter to the pixel data.
Private Types
typedef std::vector< float >	KernelValues
	Kernel values of convolution filters are stored in vectors of floats.
Private Member Functions
void	copy (const VisualConvolutionFilter &other)
	Copy method for assignment operator and copy constructor.
uint32	clamp (uint32 inVal) const
	Clamps the incoming value between 0 and 255.
Private Attributes
KernelValues	kernelValues
	Kernel values.
uint8	numberOfKernelValueColumns
	The number of columns of the kernel values.
uint8	numberOfKernelValueRows
	The number of rows of the kernel values.
float	postConvolutionScaleFactor
	The scaling of values applied after the convolution.
float	postConvolutionBias
	The bias of values applied after the convolution.
EdgeAction	edgeAction
	How the convolution filter treats edge values.

---

Detailed Description

Convolution filter for bitmap image effects.

A.k.a. kernel filter or matrix. A.k.a. correlation, impulse response, or point spread function.

---

Member Typedef Documentation

typedef std::vector<float> VizKit::VisualConvolutionFilter::KernelValues [private]

Kernel values of convolution filters are stored in vectors of floats.

---

Member Enumeration Documentation

enum VizKit::VisualConvolutionFilter::EdgeAction

How to calculate edge values.

Enumerator:

kZeroEdges
kClampEdges	< Edge values are blank.
kWrapEdges	< Edge values are clamped. < Edge values are wrapped around.

enum VizKit::VisualConvolutionFilter::Effect

Constants for various convolution filters.

Enumerator:

kNone
kBlur
kMotionBlur	< Blur filter (lowpass filter). Eliminates changes that are only a pixel wide. Rapid change from pixel to pixel is blocked while slow change is passed.
kGaussianBlur	< Motion blur filter.
kLaplacian	< Gaussian blur filter.
kEdgeDetect	< Laplacian filter.
kSharpen	< Filter for edge detection.
kEmboss	< Sharpens image (highpass filter). Rapid change from pixel to pixel is passed while slow change is blocked. < Emboss filter that gives the image a relief look (useful when bumpmapping).

---

Constructor & Destructor Documentation

VisualConvolutionFilter::VisualConvolutionFilter	(	const Effect	effect,
		const uint8	colCount = 3,
		const uint8	rowCount = 3
	)

The constructor.

Parameters:

effect	The name of the convolution filter.
colCount	The number of columns of the kernel matrix. Default = 3.
rowCount	The number of rows of the kernel matrix. Default = 3.

Remarks:

The size of the filter has to be uneven, so that it has a center (e.g. 3x3, 5x5, or 7x7).

VisualConvolutionFilter::~VisualConvolutionFilter

(

)

The destructor.

VisualConvolutionFilter::VisualConvolutionFilter

(

const VisualConvolutionFilter &

other

)

Copy constructor.

Parameters:

other

Reference to another VisualConvolutionFilter.

---

Member Function Documentation

void VisualConvolutionFilter::applyToPixelData	(	const PixelColor *	pixelData,
		uint32	imageWidth,
		uint32	imageHeight,
		int	pixelFormat,
		int	pixelDataType,
		PixelColor **	outPixelData,
		uint16	numberOfInterations = 1
	)		const

Applies the convolution filter to the pixel data.

Parameters:

pixelData	The pixel data on which the convolution filter is about to be applied.
imageWidth	The width of the image.
imageHeight	The height of the image.
pixelFormat	The format of the pixel data (rgba, luminance, etc.).
pixelDataType	The data type of the pixel data (8-bit byte, float, etc.).
outPixelData	The result of the convolution operation.
numberOfInterations	Optional number of iterations (default = 1).

uint32 VisualConvolutionFilter::clamp

(

uint32

inVal

)

const [private]

Clamps the incoming value between 0 and 255.

void VisualConvolutionFilter::copy

(

const VisualConvolutionFilter &

other

)

[private]

Copy method for assignment operator and copy constructor.

Parameters:

other

Another VisualConvolutionFilter.

void VisualConvolutionFilter::copyKernelValues

(

float *

buffer

)

const

Copies the kernel values of the convolution filter into the provided buffer.

Parameters:

[out]

buffer

The buffer into which the kernel values are written.

Remarks:

The caller has to make sure that the provided buffer is big enough to store columns * rows * sizeof(float) values.

uint8 VisualConvolutionFilter::getNumberOfKernelValueColumns

(

void

)

const

Returns the number of kernel value columns.

Returns:

The number of kernel value columns.

uint8 VisualConvolutionFilter::getNumberOfKernelValueRows

(

void

)

const

Returns the number of kernel value rows.

Returns:

The number of kernel value rows.

float VisualConvolutionFilter::getPostConvolutionBias

(

void

)

const

Returns the post convolution bias.

Returns:

The post convolution bias.

Remarks:

The bias is a coefficient that is added to each pixel value.

float VisualConvolutionFilter::getPostConvolutionScaleFactor

(

void

)

const

Returns the post convolution scale factor.

Returns:

The post convolution scale factor.

Remarks:

The scale factor is the factor with which each pixel is multiplied.

VisualConvolutionFilter & VisualConvolutionFilter::operator=

(

const VisualConvolutionFilter &

other

)

Assignment operator.

Parameters:

other

Reference to another VisualConvolutionFilter.

---

Member Data Documentation

EdgeAction VizKit::VisualConvolutionFilter::edgeAction [private]

How the convolution filter treats edge values.

KernelValues VizKit::VisualConvolutionFilter::kernelValues [private]

Kernel values.

uint8 VizKit::VisualConvolutionFilter::numberOfKernelValueColumns [private]

The number of columns of the kernel values.

uint8 VizKit::VisualConvolutionFilter::numberOfKernelValueRows [private]

The number of rows of the kernel values.

float VizKit::VisualConvolutionFilter::postConvolutionBias [private]

The bias of values applied after the convolution.

float VizKit::VisualConvolutionFilter::postConvolutionScaleFactor [private]

The scaling of values applied after the convolution.

---

The documentation for this class was generated from the following files:

• vizkit/source/VisualConvolutionFilter.h
• vizkit/source/VisualConvolutionFilter.cpp