RETRO commands


PROGRAM NAME

vmarr - Create a Marr Type Edge Detection Filter Kernel (K1)

DESCRIPTION

creates a convolution filter kernel using the following equation. pixel = [-1/(PI*var*var)]*[1-r*r/(2*var)]*exp[-r*r/(2*var)], where r is the calculated distance from the center of the kernel.

A Marr filter is the Laplacian of a Gaussian, and is intended for high lighting image edges and at the same time smooth out the noise. vmarr creates image data of type float.

REQUIRED ARGUMENTS

-o
type: outfile
desc: Marr kernel output image

OPTIONAL ARGUMENTS

-trigger
type: infile
desc: trigger input
default: {none}
-r
type: integer
desc: specify the number of rows in the kernel
default: 3
bounds: value > 0
-c
type: integer
desc: specify the number of columns in the kernel
default: 3
bounds: value > 0
-v
type: float
desc: variance of the Gaussian
default: 1
bounds: value > 0.0

EXAMPLES

SEE ALSO

kasc2val, kimportasc

RESTRICTIONS

only generates float kernels, kconvert can change the image to other data types.

REFERENCES

COPYRIGHT

Copyright (C) 1993 - 1997, Khoral Research, Inc. ("KRI") All rights reserved.