mehta-lab / recOrder

3D quantitative label-free imaging with phase and polarization
BSD 3-Clause "New" or "Revised" License
11 stars 4 forks source link

test and refine calibration assessment #55

Closed mattersoflight closed 2 years ago

mattersoflight commented 2 years ago

This issue is to document the calibration assessment that the plugin provides, and refine those assessments.

2022/01/27

Extinction was poor, because condenser was not adjusted for Kolher illumination, but the plugin suggested turning the polarizer.

Before Kohler image

After Kholer image

Update above status message to say, "Poor extinction. Possibilities are: a) polarizer and LC are not oriented properly, b) circular analyzer has wrong handedness, c) the condenser is not setup for Kohler illumination, d) a component, such as autofocus dichroic, distorts the polarization state."

camFoltz commented 2 years ago

Thanks for bringing this up. I agree that we need to be more general in the assessments.

I would also like to discuss what may be the best bounds for certain error messages. Here is the current implementation:

Note: all LC values are for the extinction state

0.2 < LC-A < 0.4 
0.4 < LC-B < 0.75
extinction >= 100
--------------------> Assess: Successful Calibration

0.2 < LC-A < 0.4
0.4 < LC-B < 0.75
80 <= extinction <= 100
--------------------> Assess: Successful Calibration, Okay Extinction Ratio

0.2 < LC-A < 0.4
0.4 < LC-B < 0.75
extinction < 90
--------------------> Assess: Poor Extinction try tunning ...

0.2 < LC-A < 0.4
 LC-B < 0.4 or LCB > 0.75
--------------------> Assess: Calibration Failed, wrong handedness of analyzer or linear polarizer 90 degrees off

LC-A < 0.2 or LCA-A > 0.4
--------------------> Assess: Calibration Failed, unknown origin of issue

Right now these values are quite arbitrary (the value of 80 for the extinction ratio for instance) and as @ieivanov mentioned in the lab meeting, if the calibration curves are off then the calibration could be successful but the values could still hit outside of this range.

camFoltz commented 2 years ago

This has been simplified in #31 to just use the extinction ratio and not any LC-values to assess the calibration. Gives ideas to the user on how to fix the problem if calibration is poor.