<table cellspacing="0" cellpadding="0" border="0" ><tr><td valign="top" style="font: inherit;"><span class="Apple-style-span" style="font-family: 'Times New Roman'; font-size: medium; "><pre>Dear Juan,</pre><pre><br></pre><pre>Level set segmentation filers have a reverse expansion flag which </pre><pre><br></pre><pre>can be turned off or on for changing the direction of the surface</pre><pre><br></pre><pre>expansion. You can change the direction of the expansion after n </pre><pre><br></pre><pre>iterations and go in the reverse direction for m iterations and </pre><pre><br></pre><pre>maximum number of iterations needed for both steps would be n+m. </pre><pre><br></pre><pre>n indicates the number of iterations needed for reaching the </pre><pre><br></pre><pre>optimal level set and m is the number of iterations needed for </pre><pre><br></pre><pre>inflating or deflating the surface.</pre><pre><br></pre><pre>Hope this
helps,</pre><pre><br></pre><pre>Dawood</pre><pre><br></pre><pre>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<</pre><pre><br></pre><pre><span class="Apple-style-span" style="font-family: 'Times New Roman'; white-space: normal; "><pre>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<</pre></span></pre><pre><br></pre><pre><br></pre><pre>>><i> On Thu, Feb 3, 2011 at 12:05 PM, Juan Cardelino
</i>>><i> <<a href="http://www.itk.org/mailman/listinfo/insight-users">juan.cardelino at gmail.com</a>> wrote:
</i>>>><i> I want to perform a very simple task: I have a curve represented as
</i>>>><i> the level set of an implicit function (coming from a level set
</i>>>><i> segmentation algorithm) and I want to inflate (or deflate) it a
</i>>>><i> certain amoun. Lets say 10 pixels from the original location. As I'm
</i>>>><i> used to work with full level sets, I usually substract 10 pixels from
</i>>>><i> the implicit function. For that you need the full implicit function.
</i>>>><i> As in ITK we use only a band near the level set, that won't work for
</i>>>><i> sure.
</i>>>><i> All the alternatives I'm thinking seem to be somehow complicated:
</i>>>><i>
</i>>>><i> 1) compute the zero contour and use fastmarching (signed) on the it,
</i>>>><i> then cut this distance function at the desired level.
</i>>>><i> 2) evolve the level set function with a constant curvature dependent
</i>>>><i> speed, until it advances the desired amount of pixels
</i>>>><i>
</i>>>><i> I'm just wondering if there is anything already built in ITK.
</i>>>><i> Thanks in advance.
</i>>>><i> Regards,
</i>>>><i> Juan</i></pre></span></td></tr></table><br>