Segmentation and edge detection

Description

Fucntions to segment images and detect edges.

Usage

    edge(x, r=0, ...)
    segment(x, cl=10, s=1.5, ...)
    fill(x, col, xoff, yoff, fuzz=10, method="floodfill", ...) ## FIXME: still not implemented, unclear API

Arguments

x

An object of Image-class.

r

The radius of the pixel neighbourhood to take into account. The 0 value enables automatic radius selection.

cl

Minimum cluster size in pixels .

s

The smoothing threshold.

col

The fill colour in any format (character, gray value or integer byte-coded value).

xoff, yoff

Point to start fill at.

fuzz

Colour tolerance, within this interval the colours are considered the same.

...

Reserved for future use.

Value

A transformed image in an object of Image-class.

Details

edge returns an image of edges between different colours in the original image. Most effective on binary images.

segment segment an image by analyzing the histograms of the color components and identifying units that are homogeneous with the fuzzy C-means technique (source and implementation ImageMagick). The smoothing threshold eliminates noise in the second derivative of the histogram. As the value is increased, you can expect a smoother second derivative.

fill should allow colour fill, but ImageMagick API is not clear and the current implementation does not seem to produce any reasonable result. Therefore the function will currently produce an error message, should be corrected in future releases.

Author(s)

Copyright (c) 2005-2006 Oleg Sklyar : osklyar@ebi.ac.uk

References

ImageMagick: http://www.imagemagick.org.

See Also

Image-class