Thank you for the response Matt, <div><br></div><div>I am able to obtain better results for my </div><div>problem on synthetic data with equal spacing.</div><div><br></div><div><meta http-equiv="content-type" content="text/html; charset=utf-8"><div>
<span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">This filter has a function </span><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">SetUseImageSpacing. </span><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">What I am wondering </span></div>
<meta http-equiv="content-type" content="text/html; charset=utf-8"><div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">is the relationship between </span><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">the spacing and the sigma value I </span></div>
<div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">should be providing. The other parameters are making perfect </span></div><div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">sense to me, as I decrease </span><font class="Apple-style-span" face="'Lucida Grande', Verdana, Geneva, Arial, sans-serif"><span class="Apple-style-span" style="font-size: 12px; ">the maximum error parameter </span></font><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">the </span></div>
<div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">kernel width should going </span><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">to increase.</span></div>
</div><div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; "><br></span></div><div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">As an example: </span></div>
<div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">If for my synthetic data with equal spacing (1,1,1), I use sigma=3,</span></div><div>
<span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">should I be using sigma=1.5 for a spacing of (0.5,0.5,0.5) with</span></div><div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">UseImageSpacing turned on? It looks like the filter internally </span></div>
<div><span class="Apple-style-span" style="font-family: 'Lucida Grande', Verdana, Geneva, Arial, sans-serif; font-size: 12px; ">divides the spacing with the spacing when this flag is on. </span></div><div><font class="Apple-style-span" face="'Lucida Grande', Verdana, Geneva, Arial, sans-serif"><span class="Apple-style-span" style="font-size: 12px;">I would appreciate insights to this function. </span></font></div>
<div><font class="Apple-style-span" face="'Lucida Grande', Verdana, Geneva, Arial, sans-serif"><span class="Apple-style-span" style="font-size: 12px;"><br></span></font></div><div><font class="Apple-style-span" face="'Lucida Grande', Verdana, Geneva, Arial, sans-serif"><span class="Apple-style-span" style="font-size: 12px;">Kind regards, </span></font></div>
<div><font class="Apple-style-span" face="'Lucida Grande', Verdana, Geneva, Arial, sans-serif"><span class="Apple-style-span" style="font-size: 12px;">Cagatay</span></font></div><div><font class="Apple-style-span" face="'Lucida Grande', Verdana, Geneva, Arial, sans-serif"><span class="Apple-style-span" style="font-size: 12px;"><br>
</span></font></div><div><div class="gmail_quote">On Wed, Sep 28, 2011 at 11:47 AM, Matt McCormick <span dir="ltr"><<a href="mailto:matt.mccormick@kitware.com">matt.mccormick@kitware.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Hi Cagatay,<br>
<div class="im"><br>
> My question:<br>
> Is there a filter in ITK that performs the traditional and more accurate<br>
> convolution ?<br>
<br>
</div><a href="http://www.itk.org/Doxygen/html/classitk_1_1DiscreteGaussianDerivativeImageFilter.html" target="_blank">http://www.itk.org/Doxygen/html/classitk_1_1DiscreteGaussianDerivativeImageFilter.html</a><br>
<div class="im"><br>
I am doing some<br>
> curvature calculation in 3D and need to be as accurate as possible. Also,<br>
> how do I deal with<br>
> different z spacing using this class or other gradient filters ?<br>
<br>
</div>This is difficult to do without resampling because the itk Image<br>
assumes uniform spacing.<br>
<font color="#888888"><br>
Matt<br>
</font></blockquote></div><br></div>