STIR
STIR copied to clipboard
add HU->mu slope conversion capability for more than 2 slopes
@KrisThielemans there are some in Duetto/PetToolbox, you mean those, or some others?
yes, the ones in the pettoolbox. 3 pieces I believe, but I don't think the 3rd matters in practice (without contrast)
@KrisThielemans can I just take them from there and add them myself? Or is it private info?
according to the letter signed by GE, we can incorporate knowledge from the pettoolbox, so this should be alright.
This is the non-contrast part of the info in there. I can find some time to add this to the json.
#Conversion table for 140 KV and Non-Contrast
KV: 140
Application: Non-Contrast
NumberOfSegments: 5
# Threshold Intercept Slope
-1000 0.0 0.0
50 0.0960 9.600E-05
1100 0.0980 5.600E-05
30000 0.1519 7.000E-06
30000 0.3619 0.000E+00
#Conversion table for 120 KV and Non-Contrast
KV: 120
Application: Non-Contrast
NumberOfSegments: 5
# Threshold Intercept Slope
-1000 0.0 0.0
50 0.0960 9.600E-05
1201 0.0982 5.110E-05
30000 0.1519 6.388E-06
30000 0.3436 0.000E+00
#Conversion table for 100 KV and Non-Contrast
KV: 100
Application: Non-Contrast
NumberOfSegments: 5
# Threshold Intercept Slope
-1000 0.0 0.0
50 0.0960 9.600E-05
1339 0.0985 4.560E-05
30000 0.1520 5.700E-06
30000 0.3230 0.000E+00
#Conversion table for 80 KV and Non-Contrast
KV: 80
Application: Non-Contrast
NumberOfSegments: 5
# Threshold Intercept Slope
-1000 0.0 0.0
50 0.0960 9.600E-05
1581 0.0989 3.840E-05
30000 0.1520 4.800E-06
30000 0.2960 0.000E+00
Seems that for now, STIR only accepts 2 slopes and 1 breakpoint. What do we do about this?
https://github.com/UCL/STIR/blob/master/src/buildblock/HUToMuImageProcessor.cxx#L171
Yes. Let's just take the first 2 pieces for now
Please put it on release4
This has been done now. I'll leave the issue open (but reassign) for later to add more then 2 slopes.