lprefilt.c File Reference

Prefiltering for linear interpolation. More...

#include <stdlib.h>
#include <math.h>
#include <fftw3.h>
#include "adaptlob.h"
#include "lprefilt.h"

Include dependency graph for lprefilt.c:

Go to the source code of this file.

Data Structures
struct	convolutionparams

Functions
void	PrefilterImage (float *Data, int Width, int Height, int NumChannels, const float *alpha, int NumFilterPairs, float ConstantFactor, boundaryhandling Boundary)
	Apply a cascade of first-order recursive filter pairs to an image. More...
void	PsfConvCoeff (float *Coeff, int NumCoeffs, float(*Psf)(float, const void *), const void *PsfParams, float(*Kernel)(float), float KernelRadius, int KernelNormalize)
	Compute PSF*kernel convolution coefficients. More...
int	PsfPreFilter (float *Data, int Width, int Height, int NumChannels, const float *Coeff, int NumCoeffs, boundaryhandling Boundary)
	Prefiltering to adjust interpolation for PSF. More...

Detailed Description

Prefiltering for linear interpolation.

Author

Pascal Getreuer getre.nosp@m.uer@.nosp@m.gmail.nosp@m..com

This file implements the prefiltering PrefilterImage on the input image needed before performing interpolation needed with B-splines and o-Moms. Also implemented here is PsfPreFilter, the prefiltering adjustment needed when interpolating with a PSF.

Copyright (c) 2010-2011, Pascal Getreuer All rights reserved.

This program is free software: you can use, modify and/or redistribute it under the terms of the simplified BSD License. You should have received a copy of this license along this program. If not, see http://www.opensource.org/licenses/bsd-license.html.

Definition in file lprefilt.c.

Function Documentation

void PrefilterImage	(	float *	Data,
		int	Width,
		int	Height,
		int	NumChannels,
		const float *	alpha,
		int	NumFilterPairs,
		float	ConstantFactor,
		boundaryhandling	Boundary
	)

Apply a cascade of first-order recursive filter pairs to an image.

Parameters

Data	the image data
Width,Height,NumChannels	image dimensions
alpha	array of alpha values
NumFilterPairs	the number of filter pairs
ConstantFactor	constant multiplicative factor to apply
Boundary	the kind of boundary handling to use

Definition at line 132 of file lprefilt.c.

Here is the caller graph for this function:

void PsfConvCoeff	(	float *	Coeff,
		int	NumCoeffs,
		float(*)(float, const void *)	Psf,
		const void *	PsfParams,
		float(*)(float)	Kernel,
		float	KernelRadius,
		int	KernelNormalize
	)

Compute PSF*kernel convolution coefficients.

Parameters

Coeff	pointer to coefficients destination array
NumCoeffs	number of coefficients to compute
Psf	the point-spread function (PSF)
PsfParams	parameters to pass to Psf
Kernel	the interpolation kernel or basis function to use
KernelRadius	support radius of Kernel
KernelNormalize	if nonzero, Kernel is normalized to sum to 1

The coefficients are the convolution Coeff[m] = (Psf * Kernel)(m) for m = 0, ..., NumCoeffs - 1. Integration is approximated using adaptive Gauss-Lobatto quadrature.

The result of this routine is used by PsfPreFilter.

Definition at line 214 of file lprefilt.c.

Here is the call graph for this function:

Here is the caller graph for this function:

int PsfPreFilter	(	float *	Data,
		int	Width,
		int	Height,
		int	NumChannels,
		const float *	Coeff,
		int	NumCoeffs,
		boundaryhandling	Boundary
	)

Prefiltering to adjust interpolation for PSF.

Parameters

Data	pointer to image data to be filtered
Width,Height,NumChannels	image dimensions
Coeff	the PSF adjustment coefficients from PsfConvCoeff
NumCoeffs	number of coefficients
Boundary	boundary handling

Convolves Data with the inverse of Coeff[0] + Coeff[1] (z + z^-1) + Coeff[2] (z^2 + z^-2) + ... via the DCT transform. Boundary handling can be either half- or whole- sample symmetric (constant extension is not supported).

Half-sample symmetric boundary handling is considerably more efficient, about 25% shorter runtime than when using whole-sample symmetry.

Definition at line 380 of file lprefilt.c.

Here is the caller graph for this function: