Thanks for the detailed/good explanation!<br><br><div class="gmail_quote">On Mon, Jun 22, 2009 at 1:06 AM, Luis Ibanez <span dir="ltr"><<a href="mailto:luis.ibanez@kitware.com">luis.ibanez@kitware.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Hi Motes,<br><br>Your new images are not appropriate input data for <br>an image registration test.<br>
<br> They contain almost no information. <br><br>The local derivatives of the metric will be zero almost <br>every where in the parametric space.<br>
<br>The cost function of the problem that you have devised<br>looks almost like a Diract delta. It has constant values <br>every where, an a very high value in a single point.<br><br>No gradient-based optimizer would figure out in what<br>
direction to walk to get to the peak of the cost function.<br><br>The only stochastic optimizer may have a chance to<br>solve the problem that you have set, and still they could<br>only solve it by accidentally hitting that singular point in<br>
which the cost function has a non trivial value.<br><br>You should use images whose spatial distribution of<br>pixel intensities can give you a local gradient on the<br>Cost Function (image metric).<br><br><br><br> Regards<br>
<br><br> Luis<br><br><br>---------------------------------------<br><div class="gmail_quote"><div><div></div><div class="h5">On Sun, Jun 21, 2009 at 3:20 PM, motes motes <span dir="ltr"><<a href="mailto:mort.motes@gmail.com" target="_blank">mort.motes@gmail.com</a>></span> wrote:<br>
</div></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div><div></div><div class="h5">I am playing around with the example from the ItkSoftwareGuide.pdf page 316. When using the brain images I get rather good results:<br>
<br><a href="http://img37.imageshack.us/img37/4316/example1z.jpg" target="_blank">http://img37.imageshack.us/img37/4316/example1z.jpg</a><br>
<br>showing in the order: fixed, moving, resampled, rescaled intensity.<br><br>I have then tried making my own very simple fixed and moving image to compare:<br><br><a href="http://img23.imageshack.us/img23/4744/example2t.jpg" target="_blank">http://img23.imageshack.us/img23/4744/example2t.jpg</a><br>
<br>Here the result are not that well (the same order as above). For some reason the box is only translated very little towards the position in the fixed image. Any ideas why it does not align better with this simple example? The resulting translation transform does not transform very correct.<br>
<br></div></div>_____________________________________<br>
Powered by <a href="http://www.kitware.com" target="_blank">www.kitware.com</a><br>
<br>
Visit other Kitware open-source projects at<br>
<a href="http://www.kitware.com/opensource/opensource.html" target="_blank">http://www.kitware.com/opensource/opensource.html</a><br>
<br>
Please keep messages on-topic and check the ITK FAQ at: <a href="http://www.itk.org/Wiki/ITK_FAQ" target="_blank">http://www.itk.org/Wiki/ITK_FAQ</a><br>
<br>
Follow this link to subscribe/unsubscribe:<br>
<a href="http://www.itk.org/mailman/listinfo/insight-users" target="_blank">http://www.itk.org/mailman/listinfo/insight-users</a><br>
<br></blockquote></div><br>
</blockquote></div><br>