<HTML><BODY style="word-wrap: break-word; -khtml-nbsp-mode: space; -khtml-line-break: after-white-space; ">Hi Jing,<DIV><BR class="khtml-block-placeholder"></DIV><DIV>actually it is quite easy to compute the eigenvectors, that you need. Since the Hessian matrix is a symmetric tensor of rank 2 (assuming, that your derivatives are continuous) you should find what you need by including these filters (and maybe some additional smoothing filters)</DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><FONT class="Apple-style-span" color="#683821" face="Monaco" size="3"><SPAN class="Apple-style-span" style="font-size: 11px;"><BR class="khtml-block-placeholder"></SPAN></FONT></DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><FONT class="Apple-style-span" color="#683821" face="Monaco" size="3"><SPAN class="Apple-style-span" style="font-size: 11px;">#include "itkSymmetricSecondRankTensor.h"</SPAN></FONT></DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><FONT class="Apple-style-span" color="#683821" face="Monaco" size="3"><SPAN class="Apple-style-span" style="font-size: 11px;">#include "itkSymmetricEigenAnalysisImageFilter.h"</SPAN></FONT></DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><FONT class="Apple-style-span" color="#683821" face="Monaco" size="3"><SPAN class="Apple-style-span" style="font-size: 11px;">#include "itkHessianRecursiveGaussianImageFilter.h"</SPAN></FONT></DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><FONT class="Apple-style-span" color="#683821" face="Monaco" size="3"><SPAN class="Apple-style-span" style="font-size: 11px;">#include "itkGradientRecursiveGaussianImageFilter.h"</SPAN></FONT></DIV><DIV><BR class="khtml-block-placeholder"></DIV><DIV>You can then easily do what you proposed.</DIV><DIV><BR class="khtml-block-placeholder"></DIV><DIV><BR class="khtml-block-placeholder"></DIV><DIV>Best regards</DIV><DIV>Ruben</DIV><DIV><BR class="khtml-block-placeholder"></DIV><DIV><BR class="khtml-block-placeholder"></DIV><DIV><BR><DIV><DIV>Am 20.06.2007 um 15:03 schrieb jing xu:</DIV><BR class="Apple-interchange-newline"><BLOCKQUOTE type="cite"><DIV>II use ITK to extract the centreline of 3D vessel ,The basic theory is to calculate the Hessian matrix of image at scale</DIV> <DIV>Then</DIV> <DIV>lamda1<lamda2<lamda3 and lamda1<lamda2<0</DIV> <DIV>The scale product of eigenvector1 and gradient of image is equal to zero </DIV> <DIV>The scale product of eigenvector2 and gradient of image is equal to zero </DIV> <DIV>But in ITK software, I only find function itkSymmetricEigenAnalysisImageFilter to deal with 3D Hessian matrix and there is no method to get the eigenvector</DIV> <DIV>I want to know there is any other fuction to get the eigenvector of 3D Hessian matrix </DIV> <DIV>Thank you </DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; ">_______________________________________________</DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; ">Insight-users mailing list</DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><A href="mailto:Insight-users@itk.org">Insight-users@itk.org</A></DIV><DIV style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0px; "><A href="http://www.itk.org/mailman/listinfo/insight-users">http://www.itk.org/mailman/listinfo/insight-users</A></DIV> </BLOCKQUOTE></DIV><BR></DIV></BODY></HTML>