Structure tensor: Difference between revisions

Revision as of 20:04, 25 September 2006

Revision as of 20:07, 25 September 2006

Line 121:

well. For example, local shifting of edge locations in minimized when

applying a [[Gaussian smoothing operator]] element-by-element to the structure

tensor <ref name="Medioni">{{cite conference|author=M. Nicolescu and G. Medioni

tensor [Nicolescu and Medioni, 2003]. Furthermore, the cancellations of

|title=Motion Segmentation with Accurate Boundaries — A Tensor Voting Approach|booktitle=Proc. IEEE Computer Vision and Pattern Recognition|volume=1|pages=pp.382–389|year=2003}}</ref>. Furthermore, the cancellations of

opposing gradient polarity directions are prevented when structure tensors

are summed <ref>{{cite journal|author=T. Brox, J. Weickert, B. Burgeth and P. Mrazek|title=Nonlinear Structure Tensors|booktitle=Universitat des Saarlandes, Tech. Report|issue=113|pages=1–32|year=2004}}</ref>.

Line 147 ⟶ 148:

The differences between the eigenvalues indicate underlying structure as

well. For example, if the value of (L1-L2)>>0, this depicts a "surfel,"

surface element, where e1 is the normal to the surface<ref>{{cite conference|authorname=M"Medioni"/>. Nicolescu and G. Medioni

|title=Motion Segmentation with Accurate Boundaries — A Tensor Voting Approach|booktitle=Proc. IEEE Computer Vision and Pattern Recognition|volume=1|pages=pp.382–389|year=2003}}</ref>.

<gallery>