question regarding fwxm calculation

Google Groups
1

Hi.

Currently I am trying to calculate the fwhm of some watertank scans with pylinac.core.profile.fwxm and I am getting not the expected results. I use the interpolate option.
When I calculate the FWHM with the fwxm I get 998.95, when I do the calculation by hand I get 1002.66.

The relevant measurement data:

index   position  values
295     -50.5     0.602185
296     -50.4     0.619593
...      max      1.222600
1298     49.8     0.614352
1299     49.9     0.597763

I tried to follow the code but it was a bit hard to read. In which way does Pylinac do the interpolation? I just did a linear interpolation between the data values.
Any hints?

With kind regards.

Thomas Bezold
2

Hi Thomas,
What height value are you using? I.e. the x parameter of the fwhm method. Default interpolation is linear but can be specified as a class attr.

3

Hi James.

I’am using the default value 50.

I input a numpy array with

x_profile = pylinac.core.profile.SingleProfile(mesvalues)

and the calculate the FWHM with:
x_profile.fwxm(interpolate=True)

4

Sorry for the late reply.

The reason is probably because pylinac takes the minimum of each side as the lower value and finds the 50% level between those values. If the lower values on each side are the same then the result will be what you expect, but if one side has some background then the result will not simply be the FWHM as you expect because pylinac tries to account for noise. This tactic is very useful for images with non-uniform background.