Fork me on GitHub
Welcome, Guest
Username: Password: Remember me

TOPIC: Computing of Eigenvectormaps

Computing of Eigenvectormaps 8 years 11 months ago #250

  • Viertann
  • Viertann's Avatar
Hi,
In the Diffusion workspace there is the possibility to compute Eigenvectormaps which are shown in grey values. I would like to know how these Maps are computed. So like in my Question to the RGB color code i need a funktional dependence of the grey value to the Eigenvectors. I hope you can help.

Thanks

Olivier
The administrator has disabled public write access.

Re: Computing of Eigenvectormaps 8 years 11 months ago #252

  • ocommowick
  • ocommowick's Avatar
  • OFFLINE
  • Administrator
  • Posts: 49
  • Thank you received: 5
  • Karma: 7
Hi again,

I suppose you're talking about the l1, l2 and l3 maps that you can compute. But I may be wrong. Technically, those values are just computed from the tensor decomposition in eigenvectors and eigenvalues. A tensor can be decomposed as T = V D Vt , where V is a matrix containing the eigenvectors and D a diagonal matrix containing the eigenvalues.

l1 is just an image of the largest value of D at each pixel. The grey value displayed is then dependent on the window/level you have set (you can see it by checking the scalar bar checkbox in the view properties toolbox). The same goes for l2 and l3 : l2 is the second largest eigenvalue at each point, l3 the smallest eigenvalue.

Is that answering your question or am I completely out of scope ?
The administrator has disabled public write access.

Re: Computing of Eigenvectormaps 8 years 11 months ago #253

  • Viertann
  • Viertann's Avatar
Hi again,
You are right, that was what I was talking about but I thought, that they would be maps of the Eigenvectors instead of Eigenvalues. What I need are the directions of the Eigenvector with the largest Eigenvalue. So that won't be the way I suppose.

Thanks for your help
The administrator has disabled public write access.
Time to create page: 0.123 seconds