saving point and polygon ROI as text files

[Kenneth Sloan]

   

>
> This must be trivial - but all I can find are complicated solutions.
>
> What I want to do:
>
> *single image
> *select and save a set of x,y coordinates (PointPicker seems to do this)
> *select and save a single polygon selection (I must be blind, I can't find this)
>
> given a good night's sleep, I'm pretty sure I could write a Java plugin to do this.
>
> Is there an easier, faster way to get there?
>
> My desired output is two files - something like .csv files like:
>
> 123,456
> 135,789
> ....
> 999,879
>
> The first file containing the points (I'm guessing the format generated by PointPicker will do.
>
> The second file containing the vertices of the polygon.
>
> I don't want to PROCESS the points or the polygon - just point&click and then save.
>
>
> REASONS: I'm creating a backup for a (so far) not quite right plugin that is supposed to FIND the points.  The images are proving to be more
> complicated that we expected, and I want a fully manual system as backup.   The rest of our pipeline is in place.  It currently expects a .csv file of the
> form:
>
> points,,,polygon
> x,y,,x,y
> 123,456,,111,222
> 135,789,,333,444
> ....
>
> I will eventually write the Java plugin to create this file directly - what I'm looking for (tonight!) is a quick and dirty basic-ImageJ point and click procedure.
>
> This SHOULD be an easy question - but I'm thrashing, so it's time to ask for assistance.

 
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Herbie-2]

Dear Kenneth Sloan,

if I understand you correctly and if you are willing to start with an
IJ-macro approach, you may have a look at the macro routine

getSelectionCoordinates(xCoordinates, yCoordinates);
"Returns two arrays containing the X and Y coordinates, in pixels, of
the points that define the current selection."

HTH

Herbie
_________________________________________
On 17.02.13 05:21, Kenneth Sloan wrote:

>
>>
>> This must be trivial - but all I can find are complicated solutions.
>>
>> What I want to do:
>>
>> *single image
>> *select and save a set of x,y coordinates (PointPicker seems to do this)
>> *select and save a single polygon selection (I must be blind, I can't find this)
>>
>> given a good night's sleep, I'm pretty sure I could write a Java plugin to do this.
>>
>> Is there an easier, faster way to get there?
>>
>> My desired output is two files - something like .csv files like:
>>
>> 123,456
>> 135,789
>> ....
>> 999,879
>>
>> The first file containing the points (I'm guessing the format generated by PointPicker will do.
>>
>> The second file containing the vertices of the polygon.
>>
>> I don't want to PROCESS the points or the polygon - just point&click and then save.
>>
>>
>> REASONS: I'm creating a backup for a (so far) not quite right plugin that is supposed to FIND the points.  The images are proving to be more
>> complicated that we expected, and I want a fully manual system as backup.   The rest of our pipeline is in place.  It currently expects a .csv file of the
>> form:
>>
>> points,,,polygon
>> x,y,,x,y
>> 123,456,,111,222
>> 135,789,,333,444
>> ....
>>
>> I will eventually write the Java plugin to create this file directly - what I'm looking for (tonight!) is a quick and dirty basic-ImageJ point and click procedure.
>>
>> This SHOULD be an easy question - but I'm thrashing, so it's time to ask for assistance.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Kenneth Sloan]

Well, I was hoping for a nice, simple sequence of "point and click" using standard tools:

a) click here
b) select n points (leaving small bread crumbs behind)
c) click here
d) type file name  (so far, I have described how PointPicker works)
e) click here
f) select n points (leaving a polygon behind)
g) double click here (now we have a closed polygon)
h) click here
i) type file name

done!

Or, perhaps I was hoping there was a "PolygonPicker" that I didn't know about.

I'm more of a Java-plugin kind of guy, rather than an IJ-macro person - a character flaw, I suppose.

I'll write the Java version to get my work done (actually....it looks like a student has just finished that - along with the
inevitable bells and whistles) and then do it as a macro, for penance.

If anyone cares - here's my application: I want to select many, many cell centers (on the order of 100-200 per field of view) AND ALSO
select a polygon surrounding a "good, error- and artifact-free area".  This information is fed into a very long pipeline of (non-IJ) programs
for analysis.  Essentially, I compute the Delaunay Triangulation and Voronoi Diagram of the points and then keep only the VD cells which are
entirely inside the "good region".  These VD cells then form the basis for further analysis (the usual spatial statistics, plus gathering of autofluorescence
data from OTHER images of the same field of view.  

The entire pipeline is in place and fully functional - using a semi-automatic method of identifying the cell centers.  We have a Java IJ-plugin that was doing a pretty good job of finding most of the cell centers and allowing the user to make minor edits (remove THESE points, add THOSE points) before saving the points and the polygon to a file.  It was working so well that we were contemplating the task of extending it to find the cell boundaries (which are similar to, but not exactly the same as, the Voronoi Diagram of the cell centers).  This looked challenging, but do-able.  

Alas, the semi-automatic method has just hit a speed bump (the latest images are too challenging), and I wanted a quick-and-dirty way to MANUALLY perform the first step in the pipeline, without devoting a lot of effort.  I found PointPicker, which is EXACTLY what I want for the cell centers - but was stumped by the task of selecting and saving the polygon surrounding the "good" region.

Hmmm...it occurs to me that "PointPicker" might be modified to work in "polygon" mode.  But, that's probably just a bit more work than this project is worth.
Allow me to suggest it to the keepers of "PointPicker" (and please tell me if you do it!)

--
Kenneth Sloan
[hidden email]


You begin to see why I was interested in "quick-and-dirty"?

--
Kenneth Sloan
[hidden email]



On Feb 17, 2013, at 07:05 , Herbie wrote:

> Dear Kenneth Sloan,
>
> if I understand you correctly and if you are willing to start with an IJ-macro approach, you may have a look at the macro routine
>
> getSelectionCoordinates(xCoordinates, yCoordinates);
> "Returns two arrays containing the X and Y coordinates, in pixels, of the points that define the current selection."
>
> HTH
>
> Herbie
> _________________________________________
> On 17.02.13 05:21, Kenneth Sloan wrote:
>>
>>>
>>> This must be trivial - but all I can find are complicated solutions.
>>>
>>> What I want to do:
>>>
>>> *single image
>>> *select and save a set of x,y coordinates (PointPicker seems to do this)
>>> *select and save a single polygon selection (I must be blind, I can't find this)
>>>
>>> given a good night's sleep, I'm pretty sure I could write a Java plugin to do this.
>>>
>>> Is there an easier, faster way to get there?
>>>
>>> My desired output is two files - something like .csv files like:
>>>
>>> 123,456
>>> 135,789
>>> ....
>>> 999,879
>>>
>>> The first file containing the points (I'm guessing the format generated by PointPicker will do.
>>>
>>> The second file containing the vertices of the polygon.
>>>
>>> I don't want to PROCESS the points or the polygon - just point&click and then save.
>>>
>>>
>>> REASONS: I'm creating a backup for a (so far) not quite right plugin that is supposed to FIND the points.  The images are proving to be more
>>> complicated that we expected, and I want a fully manual system as backup.   The rest of our pipeline is in place.  It currently expects a .csv file of the
>>> form:
>>>
>>> points,,,polygon
>>> x,y,,x,y
>>> 123,456,,111,222
>>> 135,789,,333,444
>>> ....
>>>
>>> I will eventually write the Java plugin to create this file directly - what I'm looking for (tonight!) is a quick and dirty basic-ImageJ point and click procedure.
>>>
>>> This SHOULD be an easy question - but I'm thrashing, so it's time to ask for assistance.
>>
>> --
>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html