Image Processing Reference
In-Depth Information
Fig. 14.12
Example tensors and their corresponding vector field patterns [
28
]. © IEEE Reprinted,
with permission, from IEEE Transactions on Visualization and Computer Graphics 15(1)
Fig. 14.13
The tensor decomposition in Eq. (
14.4
) can be adapted to symmetric tensors. In this
example the symmetric tensor is the curvature tensor in the surface. Note that this tensor decom-
position can lead to surface classification and feature extraction [
15
]. © IEEE Reprinted, with
permission, from IEEE Transactions on Visualization and Computer Graphics 18(6)
The decomposition in Eq. (
14.4
) can also be used to
symmetric
tensor fields. In
this case
γ
r
=
0 and the tensor can be rewritten as:
10
01
cos
θ
sin
θ
ρ
φ
+
ρ
φ
sin
cos
(14.5)
sin
θ
−
cos
θ
2
d
where
s
again is the tensor magnitude. Like Eq. (
14.4
), this equation
is a special case of Eq. (
14.2
) where one of the three components disappears (the
anti-symmetric component). Nieser et al. [
15
] have applied this to the curvature
tensor to extract surface features for remeshing purposes. Figure
14.13
illustrates this
ρ
=
γ
+
γ
Search WWH ::
Custom Search