I’m trying to use the example from Register Image to Another Using Landmarks — v5.3.0 to set up a registration based on matching set of points. The estimated transform seems to me OK:
AffineTransform (0x157604c30)
RTTI typeinfo: itk::AffineTransform<double, 2u>
Reference Count: 3
Modified Time: 2361
Debug: Off
Object Name:
Observers:
none
Matrix:
0.999999 0.000201643
-0.000235857 0.999915
Offset: [4.35633, 3.04601]
Center: [0, 0]
Translation: [4.35633, 3.04601]
Inverse:
1 -0.00020166
0.000235878 1.00009
Singular: 0
Unfortunately, it looks like the resulting image has a higher translation than this, and the image is also brighter, which doesn’t make sense…
The two images are from the same set and would have the same spacing and stuff.
Any idea what I’m missing?