// ==========================================================
// Upsampling / downsampling filters
//
// Design and implementation by
// - Hervé Drolon (drolon@infonie.fr)
//
// This file is part of FreeImage 3
//
// COVERED CODE IS PROVIDED UNDER THIS LICENSE ON AN "AS IS" BASIS, WITHOUT WARRANTY
// OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES
// THAT THE COVERED CODE IS FREE OF DEFECTS, MERCHANTABLE, FIT FOR A PARTICULAR PURPOSE
// OR NON-INFRINGING. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE COVERED
// CODE IS WITH YOU. SHOULD ANY COVERED CODE PROVE DEFECTIVE IN ANY RESPECT, YOU (NOT
// THE INITIAL DEVELOPER OR ANY OTHER CONTRIBUTOR) ASSUME THE COST OF ANY NECESSARY
// SERVICING, REPAIR OR CORRECTION. THIS DISCLAIMER OF WARRANTY CONSTITUTES AN ESSENTIAL
// PART OF THIS LICENSE. NO USE OF ANY COVERED CODE IS AUTHORIZED HEREUNDER EXCEPT UNDER
// THIS DISCLAIMER.
//
// Use at your own risk!
// ==========================================================
#ifndef _FILTERS_H_
#define _FILTERS_H_
/**
CGenericFilter is a generic abstract filter class used to access to the filter library.
Filters used in this library have been mainly taken from the following references :
Main reference :
Paul Heckbert, C code to zoom raster images up or down, with nice filtering.
UC Berkeley, August 1989. [online] http://www-2.cs.cmu.edu/afs/cs.cmu.edu/Web/People/ph/heckbert.html
Heckbert references :
- Oppenheim A.V., Schafer R.W., Digital Signal Processing, Prentice-Hall, 1975
- Hamming R.W., Digital Filters, Prentice-Hall, Englewood Cliffs, NJ, 1983
- Pratt W.K., Digital Image Processing, John Wiley and Sons, 1978
- Hou H.S., Andrews H.C., "Cubic Splines for Image Interpolation and Digital Filtering",
IEEE Trans. Acoustics, Speech, and Signal Proc., vol. ASSP-26, no. 6, pp. 508-517, Dec. 1978.
*/
class CGenericFilter
{
protected:
#define FILTER_PI double (3.1415926535897932384626433832795)
#define FILTER_2PI double (2.0 * 3.1415926535897932384626433832795)
#define FILTER_4PI double (4.0 * 3.1415926535897932384626433832795)
/// Filter support
double m_dWidth;
public:
/// Constructor
CGenericFilter (double dWidth) : m_dWidth (dWidth) {}
/// Destructor
virtual ~CGenericFilter() {}
/// Returns the filter support
double GetWidth() { return m_dWidth; }
/// Change the filter suport
void SetWidth (double dWidth) { m_dWidth = dWidth; }
/// Returns F(dVal) where F is the filter's impulse response
virtual double Filter (double dVal) = 0;
};
// -----------------------------------------------------------------------------------
// Filters library
// All filters are centered on 0
// -----------------------------------------------------------------------------------
/**
Box filter
Box, pulse, Fourier window, 1st order (constant) b-spline.
Reference :
Glassner A.S., Principles of digital image synthesis. Morgan Kaufmann Publishers, Inc, San Francisco, Vol. 2, 1995
*/
class CBoxFilter : public CGenericFilter
{
public:
/**
Constructor
Default fixed width = 0.5
*/
CBoxFilter() : CGenericFilter(0.5) {}
virtual ~CBoxFilter() {}
double Filter (double dVal) { return (fabs(dVal) <= m_dWidth ? 1.0 : 0.0); }
};
/** Bilinear filter
*/
class CBilinearFilter : public CGenericFilter
{
public:
CBilinearFilter () : CGenericFilter(1) {}
virtual ~CBilinearFilter() {}
double Filter (double dVal) {
dVal = fabs(dVal);
return (dVal < m_dWidth ? m_dWidth - dVal : 0.0);
}
};
/**
Mitchell & Netravali's two-param cubic filter
The parameters b and c can be used to adjust the properties of the cubic.
They are sometimes referred to as "blurring" and "ringing" respectively.
The default is b = 1/3 and c = 1/3, which were the values recommended by
Mitchell and Netravali as yielding the most visually pleasing results in subjective tests of human beings.
Larger values of b and c can produce interesting op-art effects--for example, try b = 0 and c = -5.
Reference :
Don P. Mitchell and Arun N. Netravali, Reconstruction filters in computer graphics.
In John Dill, editor, Computer Graphics (SIGGRAPH '88 Proceedings), Vol. 22, No. 4, August 1988, pp. 221-228.
*/
class CBicubicFilter : public CGenericFilter
{
protected:
// data for parameterized Mitchell filter
double p0, p2, p3;
double q0, q1, q2, q3;
public:
/**
Constructor
Default fixed width = 2
@param b Filter parameter (default value is 1/3)
@param c Filter parameter (default value is 1/3)
*/
CBicubicFilter (double b = (1/(double)3), double c = (1/(double)3)) : CGenericFilter(2) {
p0 = (6 - 2*b) / 6;
p2 = (-18 + 12*b + 6*c) / 6;
p3 = (12 - 9*b - 6*c) / 6;
q0 = (8*b + 24*c) / 6;
q1 = (-12*b - 48*c) / 6;
q2 = (6*b + 30*c) / 6;
q3 = (-b - 6*c) / 6;
}
virtual ~CBicubicFilter() {}
double Filter(double dVal) {
dVal = fabs(dVal);
if(dVal < 1)
return (p0 + dVal*dVal*(p2 + dVal*p3));
if(dVal < 2)
return (q0 + dVal*(q1 + dVal*(q2 + dVal*q3)));
return 0;
}
};
/**
Catmull-Rom spline, Overhauser spline
When using CBicubicFilter filters, you have to set parameters b and c such that
b + 2 * c = 1
in order to use the numerically most accurate filter.
This gives for b = 0 the maximum value for c = 0.5, which is the Catmull-Rom
spline and a good suggestion for sharpness.
References :
- Mitchell Don P., Netravali Arun N., Reconstruction filters in computer graphics.
In John Dill, editor, Computer Graphics (SIGGRAPH '88 Proceedings), Vol. 22, No. 4, August 1988, pp. 221-228.
- Keys R.G., Cubic Convolution Interpolation for Digital Image Processing.
IEEE Trans. Acoustics, Speech, and Signal Processing, vol. 29, no. 6, pp. 1153-1160, Dec. 1981.
*/
class CCatmullRomFilter : public CGenericFilter
{
public:
/**
Constructor
Default fixed width = 2
*/
CCatmullRomFilter() : CGenericFilter(2) {}
virtual ~CCatmullRomFilter() {}
double Filter(double dVal) {
if(dVal < -2) return 0;
if(dVal < -1) return (0.5*(4 + dVal*(8 + dVal*(5 + dVal))));
if(dVal < 0) return (0.5*(2 + dVal*dVal*(-5 - 3*dVal)));
if(dVal < 1) return (0.5*(2 + dVal*dVal*(-5 + 3*dVal)));
if(dVal < 2) return (0.5*(4 + dVal*(-8 + dVal*(5 - dVal))));
return 0;
}
};
/**
Lanczos-windowed sinc filter
Lanczos3 filter is an alternative to CBicubicFilter with high values of c about 0.6 ... 0.75
which produces quite strong sharpening. It usually offers better quality (fewer artifacts) and a sharp image.
*/
class CLanczos3Filter : public CGenericFilter
{
public:
/**
Constructor
Default fixed width = 3
*/
CLanczos3Filter() : CGenericFilter(3) {}
virtual ~CLanczos3Filter() {}
double Filter(double dVal) {
dVal = fabs(dVal);
if(dVal < m_dWidth) {
return (sinc(dVal) * sinc(dVal / m_dWidth));
}
return 0;
}
private:
double sinc(double value) {
if(value != 0) {
value *= FILTER_PI;
return (sin(value) / value);
}
return 1;
}
};
/**
4th order (cubic) b-spline
*/
class CBSplineFilter : public CGenericFilter
{
public:
/**
Constructor
Default fixed width = 2
*/
CBSplineFilter() : CGenericFilter(2) {}
virtual ~CBSplineFilter() {}
double Filter(double dVal) {
dVal = fabs(dVal);
if(dVal < 1) return (4 + dVal*dVal*(-6 + 3*dVal)) / 6;
if(dVal < 2) {
double t = 2 - dVal;
return (t*t*t / 6);
}
return 0;
}
};
// -----------------------------------------------------------------------------------
// Window function library
// -----------------------------------------------------------------------------------
/**
Blackman window
*/
class CBlackmanFilter : public CGenericFilter
{
public:
/**
Constructor
Default width = 0.5
*/
CBlackmanFilter (double dWidth = double(0.5)) : CGenericFilter(dWidth) {}
virtual ~CBlackmanFilter() {}
double Filter (double dVal) {
if(fabs (dVal) > m_dWidth) {
return 0;
}
double dN = 2 * m_dWidth + 1;
dVal /= (dN - 1);
return 0.42 + 0.5*cos(FILTER_2PI*dVal) + 0.08*cos(FILTER_4PI*dVal);
}
};
#endif // _FILTERS_H_