Convolution of an input image with an arbitrary kernel (image).
Input image:
Image to be filtered
Kernel image:
The kernel used for convolution (the given kernel is used as it is in the convolution and NOT mirrored at its origin)
Kernel normalization:
Flag to normalize the kernel image. If true the kernel image's pixel values are normalized to sum to 1.
Boundary padding:
Padding of image: Method of how to simulate pixel values outside the image domain.
PADDING_ZERO: Values outside the image domain are assumed to be zero.
PADDING_BORDER: Values outside the image domain correspond to value of nearest pixel in the image domain.
PADDING_MIRROR: Values of the image are mirrored outside of the image domain along the image border.
PADDING_PERIODIC: Values are repeated, i.e. the image is assumed to be periodical with period equal to the image dimensions (as assumed for DFT)
Kernel origin:
The coordinate of the kernel image used as origin. If not specified, the center of the kernel image rounded down is used as origin.
Result image
The filtered image of type MTBImageType.MTB_DOUBLE