Failure to generate 3D calibration

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Load the calibration stack found here:
http://www.cleterrier.net/up/Z_calibration.tif

2. Launch the 3D calibration with the parameters as specified in the 
screenshots attached  (main dialog and camera setup)

What is the expected output? What do you see instead?
The calibtation fails with the following error :

(Fiji Is Just) ImageJ 1.48r; Java 1.6.0_65 [64-bit]; Mac OS X 10.9.1; 131MB of 
6491MB (2%)

java.lang.ArrayIndexOutOfBoundsException: 0
    at cz.cuni.lf1.lge.ThunderSTORM.calibration.IterativeQuadraticFitting.fitParams(IterativeQuadraticFitting.java:71)
    at cz.cuni.lf1.lge.ThunderSTORM.calibration.CalibrationProcess.fitQuadraticPolynomials(CalibrationProcess.java:182)
    at cz.cuni.lf1.lge.ThunderSTORM.CylindricalLensCalibrationPlugin.run(CylindricalLensCalibrationPlugin.java:124)
    at ij.IJ.runUserPlugIn(IJ.java:199)
    at ij.IJ.runPlugIn(IJ.java:163)
    at ij.Executer.runCommand(Executer.java:131)
    at ij.Executer.run(Executer.java:64)
    at java.lang.Thread.run(Thread.java:695)

What version of the product are you using? On what operating system?

(Fiji Is Just) ImageJ 1.48r; Java 1.6.0_65 [64-bit]; Mac OS X 10.9.1

Please provide any additional information below.

I'm not sure about the camera setup dialog, I've entered the camera gain before 
EM gain applied (63.7 photons/AD counts), and with an EM gain of 300 the actual 
number of photons/AD counts is 0.212

Original issue reported on code.google.com by lechrist...@gmail.com on 24 Feb 2014 at 2:17

Attachments:

• Zcalib1.png
• Zcalib2.png

[GoogleCodeExporter]

Hello, thank you for the feedback. We will handle the exception and show a 
reasonable error message instead.

Solution for your problem is to make the fitting radius larger. Currently it is 
set to 2 which means that a single molecule can't be larger than 5px (2*2+1 -- 
it is a circle with radius 2 + the center pixel). When I looked at the your 
calibration data, the seem to be ~10px large when they expand to elipses. 
Setting the fitting radius to 5 (2*5+1 px) took care of the problem.

Original comment by zitmen@gmail.com on 24 Feb 2014 at 4:41

Attachments:

• cal.yaml
• screenshot-calibration.png

[GoogleCodeExporter]

Citing: "I'm not sure about the camera setup dialog, I've entered the camera 
gain before EM gain applied (63.7 photons/AD counts), and with an EM gain of 
300 the actual number of photons/AD counts is 0.212"

Response: number of photons per single AD count is not affected by EM gain

Original comment by zitmen@gmail.com on 25 Feb 2014 at 4:29

• Changed state: Started

[GoogleCodeExporter]

fixed in revision c78ca77f1f55

Original comment by zitmen@gmail.com on 25 Feb 2014 at 5:21

• Changed state: Fixed
• Added labels: Priority-Low
• Removed labels: Priority-Medium

[GoogleCodeExporter]

Citing: "number of photons per single AD count is not affected by EM gain"
Is that really so? On the Andor site it says 
signal-counts*pre-amp/(EM-gain*QE(@wavelength-of-interest)) 
[http://www.andor.com/learning-academy/count-convert-quantifying-data-in-electro
ns-and-photons] ... Am I missing something?

Thanks a lot for clarification!

Original comment by zuzu.qm...@gmail.com on 3 Mar 2014 at 4:03

[GoogleCodeExporter]

Well I guess it depends on what you call ""number of photons per single AD 
count" I guess. But I'm quite sure that the final intensity value of the pixels 
depends on the EM gain. So to convert it into photons, you have to take into 
account both the A/D gain and the EM gain. I don't agree with the use of the QE 
though: if you want to know the precision of the localization, you take into 
account the photoelectrons that have been detected, not the one that you didn't 
due to the lack of camera sensitivity. IE a camera with a very low QE will not 
result in more precision in localization than a camera with high QE, for the 
same number of photoelectrons detected.

Original comment by lechrist...@gmail.com on 3 Mar 2014 at 4:09

[GoogleCodeExporter]

"AD count" in our case is the value of "pre-amp" on the Andor site. See the 
help in the Camera setup dialog for explanation of the used variables. The 
equation is the same as on the Andor site, only we do not take QE into account.

Original comment by krize...@gmail.com on 3 Mar 2014 at 4:48