MathGroup Archive 2011

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: x y coordinates of image pixels for an image outline

  • To: mathgroup at smc.vnet.net
  • Subject: [mg118119] Re: x y coordinates of image pixels for an image outline
  • From: Stefan <wutchamacallit27 at gmail.com>
  • Date: Wed, 13 Apr 2011 05:52:56 -0400 (EDT)
  • References: <io17cv$i44$1@smc.vnet.net>

On Apr 12, 5:52 am, william parr <willpower... at hotmail.com> wrote:
> Dear Mathgroup,
>
> I think this problem is fairly straight forward, but I'm struggling with it!
>
> essentially, I want to;
> 1) determine/assign x y coordinate values to the black pixels in a binarised image.
> 2) select the outline black pixels (and get their x y coordinates) on each line of the image.
>
> This falls within a larger project of wanting to recover the outside outline (ie surface outline with x y coordinates) of a bone from a CT slice image. I would then probably aim to create a 3D point cloud for the external bone surface by assigning each slice a z coordinate value. However, unless anyone has a very clever (ie not processor intensive) way of of creating a polygon mesh from the 3D point cloud, I will probably export the point cloud for surface meshing in another program.
>
> thanks in advance,
>
> Will


Will,
  I came across a similar task not too long ago, where I wanted to
compute a list of points in the complex plane which lie on the
boundary of the Mandelbrot set. I did indeed use Binarize[] on an
image I had rendered, and then proceeded by using
MorphologicalPerimeter[], which is new and very useful and
interesting. This takes the binarized image and returns a binary image
with white pixels along the boundaries between components. You can
then use Position[] to find the image's pixel coordinates of these
points, and then transform them back to whatever actual scale/
orientation.

Binarize[image]
MorphologicalPerimeter[%]
Position[ImageData[%],1]
Map[yourCoordinateTransformation,%]
Export["vertex_coordinates.whatever",%]

-Stefan
ps. the notebook where I worked out my solution is actually on my
website, at this location (warning, lots of images on this page, the
description and .nb is at the bottom): http://people.virginia.edu/~sws4e/MathArt/Julia.html


  • Prev by Date: Re: FinancialData
  • Next by Date: Re: What are the issues with Wolfram demonstration site these days?
  • Previous by thread: Re: x y coordinates of image pixels for an image outline
  • Next by thread: Re: x y coordinates of image pixels for an image outline