convolve(6D)

NAME
     convolve - convolve an input image with a kernel

SYNOPSIS
     convolve inimage outimage kernelimage [-m -d]

DESCRIPTION
     convolve convolves the input image with a kernel image.  Kernel images
     can be created with the program greyscale.  To do a convolution, the
     kernel image is stepped across the surface of the input image.  At each
     position, all the image values under the kernel are multiplied by the
     corresponding kernel values.  The sum of all these products is divided by
     the sum of all the values in the kernel.  The result is put into the
     output image.  If the -m option is given, the maximum value of all the
     multiplies is put into the output image instead of the normalized sum of
     the products.  If the -d option is given, a delta value is calculated
     wherever the kernel is greater than 128.

SEE ALSO
     blur(6D)