Moving average

Collaboration diagram for Moving average:

Functions
int	MovingAverage (double InputData[], double OutputData[], long SignalLength, long KernelOrigin, long KernelLength, enum TBoundaryConvention Convention)
int	MovingAverageVolume (float *VolumeSource, float *VolumeDestination, long Nx, long Ny, long Nz, long KernelOrigin, long KernelLength, enum TBoundaryConvention Convention)

Detailed Description

Function Documentation

MovingAverage()

int MovingAverage	(	double	InputData[],
		double	OutputData[],
		long	SignalLength,
		long	KernelOrigin,
		long	KernelLength,
		enum TBoundaryConvention	Convention
	)

Moving average (1D). The specified boundary convention applies to the input data only, not to the kernel. The boundary convention applied to the kernel is FiniteDataSupport.

The kernel has a constant value and sums up to 1.0. The input and the output have the same length. The origin for the kernel is given with respect to the leftmost sample [0].

success: return(!ERROR); failure: return(ERROR)

General structure is as follows:

for (i = 0L; (i < SignalLength); i++) {
  Sum = 0.0;
  for (j = -Infinity; (j <= Infinity); j++)
     Sum += InputData[j] * Kernel[KernelOrigin + i - j];
  OutputData[i] = Sum;
}

MovingAverageVolume()

int MovingAverageVolume	(	float *	VolumeSource,
		float *	VolumeDestination,
		long	Nx,
		long	Ny,
		long	Nz,
		long	KernelOrigin,
		long	KernelLength,
		enum TBoundaryConvention	Convention
	)

Moving average (3D). The specified boundary convention applies to the input data only, not to the kernel. The boundary convention applied to the kernel is FiniteDataSupport. VolumeSource is a (float)volume of size (Nx x Ny x Nz). OutputData is a (float)volume of size (Nx x Ny x Nz). The origin for the kernel is given with respect to the leftmost sample [0]. The 1D kernel is applied successively to each principal direction in a separable fashion.

success: return(!ERROR); failure: return(ERROR).

General structure is as follows:

for (i = 0L; (i < SignalLength); i++) {
   Sum = 0.0;
   for (j = i - KernelOrigin; (j < (KernelLength + i - KernelOrigin)); j++)
        Sum += InputData[j];
   OutputData[i] = Sum / KernelLength;
 }