Bug 637 – Directionality tool problem. Failure to set the Histogram start option. Always result relative to 0º from original picture.

Bug 637 - Directionality tool problem. Failure to set the Histogram start option. Always result relative to 0º from original picture.


Summary:	Directionality tool problem. Failure to set the Histogram start option. Alway...

Status:	RESOLVED INVALID

Product:	Fiji
Classification:	Unclassified
Component:	Plugins
Version:	unspecified
Hardware:	Macintosh Mac OS

Importance:	P5 critical
Assigned To:	Jean-Yves Tinevez

URL:

Depends on:
Blocks:

Reported:	2013-07-18 10:27 CDT by Juan Manuel Gomez
Modified:	2014-06-16 16:41 CDT
CC List:	2 users (show)

See Also:

Attachments
Result for Directionality Analysis in which the peak should be 90ºC (13.74 KB, image/png) 2013-07-18 10:28 CDT, Juan Manuel Gomez

Description Juan Manuel Gomez 2013-07-18 10:27:48 CDT

Problem found in FIJI Directionality tool (Analysis>Directionality)

This tool allows to set a reference angle to measure the average direction of the signal in a picture. The option is "Histogram Start".

The Fiji version that was available to be downloaded in March was working OK. The problem appear afterwards with an automatic update and is also present in the newest version.

Bug Description:

The average angle returned by the Directionaly tool is always using as a reference the 0º even though I modify the option "Histogram Start" different from 0º value. No matter which is the angle for the Histogram Start always the peak of the gaussian will have the direction relative to 0ºC (the 0ºc from the picture).

Please see the attached file: That's a result in which the peak of the gaussian should be around 90ºC since I set the histogram start option to give that result. However, the histogram is still returning the peak direction from the 0ºC reference.

Please do contact me for further details. I hope I was more or less clear.

Comment 1 Juan Manuel Gomez 2013-07-18 10:28:39 CDT

Created attachment 110
Result for Directionality Analysis in which the peak should be 90ºC

Comment 2 Jean-Yves Tinevez 2013-07-18 11:07:16 CDT

Hi Juan, thanks for the report.

It seems that either I do not understand the bug description, either what you are describing is not a bug. 

The histogram start is just for plotting. It does not change the histogram content. If you have a peak at 50º, it will always be at 50º in the histogram, wherever you put its start. Well actually it may wrap by 180º but anyway. The histogram start is no a reference, it is a convenience settings to put peaks at the center of the histogram, even if they are far from 0º.

Did I get what you mean correctly?
Best
jy

Comment 3 Juan Manuel Gomez 2013-07-18 12:18:04 CDT

Dear Jean-Yves,

Thank you very much for your answer!

The problem is as you mention, for plotting but also for returning the result.

When I started using this tool, if I was changing the Histogram start value, the result was plotted relative to the angle I set up as starting point.

If you check the graphic I uploaded in my bug description, you'll see that the peak in that graph is in 160 (aprox.), however, I set up the histogram start to give that peak in 90. This means that in the plot, it should write in the x axis 90º where the peak is. However, even though I change the Histogram start value (from 0º) the peak is fixed in 160º. And that is not correct (and was not like this before). Moreover, changing the Histogram start, should return a different value in the peak. Eg: Assume for a second that the plot I uploaded is correct and the peak is in 160º (it should be written 90 insted, remember?). If know I analyse the directionality again, but setting the Histogram start as 160º the peak should be in 0º. But that is not happening. The plot will be rearranged like it would look like with that Histogram start, but the x axis will still state that the maximum is in 0.

We can skype if you want and explain you in more detail.

I hope I was clear this time!

Comment 4 Juan Manuel Gomez 2013-07-18 12:20:38 CDT

correction on my last sentence:

(...) but the x axis will still state that the maximum is in 160.

Comment 5 Jean-Yves Tinevez 2013-10-16 02:44:09 CDT

Hi Juan

I gave some thought about your report. 
And to be honest what you are reporting is not a bug, and the behavior you describe is the expected one. 

Changing the start point of the histogram simply shifts the plot. It does not change the content. For instance if you have a peak at 0º, and if you set the histogram start at -90º (the default btw), the x-axis of the histogram will range from -90º to 90º, and the peak will be displayed - as expected - at 0º.

If you now set the start to -45º, the results will be the same, but the plot will range from -45º to 135º. But of course you expect the peak to be still at 0º. 

Changing the start value just change the display of the histogram, b not its content. 

If this is really what you meant, I would like to close this issue.

best
jy

Comment 6 Curtis Rueden 2014-06-16 16:41:52 CDT

Closing due to lack of feedback.