[Insight-users] itkGradientVectorFlow remarks

[Mathieu Malaterre]

Hi all,
 I've had a look at itkGradientVectorFlow and found that:
1. Lines 230 and 231 could be placed outside of the loop (for (i=0 ...)
2. Furthermore as LaplacianFilter is used, you needn't divide again by 
m_Step[j] (remove old line 231)
3. you can take advantage of C++, that is to say += , /= ...
4. finally why not take advantage of the symmetry of the expression and 
use a kernel of (example for 2D):

[0                1                0]
[1    (-3 - b*m_TimeStep)    1]
[0                1                0]
which is a separable kernel (even if b is not const, we can use an 
address) ?

Just my 2 cents remarks,
mathieu

-- 
Le superflu n'est inutile qu'ŕ ceux qui pensent que le nécessaire est suffisant (P. Desproges)

Mathieu Malaterre
170 chemin des minimes
01600 PARCIEUX
http://mathieu.malaterre.com