Hi folks,
I have just started working with the Siena package to quantify atrophy. As a test, I've run siena on the same image, as well as a slightly modified image where I've removed some white matter.
Same image for time point 1 and time point 2:
--------------------------------------------------------
- Curiously, when Siena was run on the same image, the atrophy measurements for A to B and B to A are non-zero and the same sign (not correcting for direction) meaning A to B has growth and B to A has growth . While this results in a total atrophy of 0 ( AtoB + -1*(BtoA) ), I can't understand why a constant is being introduced when the images are the same. Does the algorithm introduce a constant offset and as a result we should only ever consider the final PBVC?
/usr/local/fsl/bin/fast B_halfwayto_A_brain > B_halfwayto_A_brain.vol 2>&1
finding brain edges
finding flow
AREA 242168 mm^2
VOLC 1064.67 mm^3
RATIO 0.00439641 mm
PBVC 0.170188 %
/usr/local/fsl/bin/fast A_halfwayto_B_brain > A_halfwayto_B_brain.vol 2>&1
finding brain edges
finding flow
AREA 242168 mm^2
VOLC 1064.67 mm^3
RATIO 0.00439641 mm
PBVC 0.170188 %
finalPBVC 0 %
Same base image for time point 1 and 2 but some white matter from 2 is manually removed:
----------------------------------------------------------------------------------------------------------------
- Results seem sensible with atrophy being recorded appropriately around the constructed lesion but the lesion does bias the registration slightly leading to edge differences.
Thanks for your help.
Amlish
|