Login  Register

Re: Pair correlation function on ImageJ

Posted by Richard Mort-2 on Jul 16, 2015; 10:09am
URL: http://imagej.273.s1.nabble.com/Pair-correlation-function-on-ImageJ-tp5013569p5013592.html

Hi Xaio and Herbie,

Apologies that this is straying in its relevance from the ImageJ list.
Please email me off list if I can help further.

But for completeness here is a method to get the g(r) for the pairwise
correlation function (images attached). You might need to consider
cropping the image depending on what you are interested in.

Herbie I'd be interested to know if something similar can be generated
using standard ImageJ methods.

Using R and ImageJ in a workflow like this can be very useful especially
when complex summary statistics are needed.

Best wishes
Richard

In ImageJ:

///////////////////////////////////
//something like this gives you the positions
run("8-bit");
setThreshold(148, 255);
run("Analyze Particles...", "size=2-Infinity display exclude clear add
in_situ");
selectWindow("Results");
saveAs("Results", "/your_dir/PPP.csv");

//end

Then in R with the package spatstat installed:

###################
library(spatstat)
setwd("/your_dir/")
data1 = read.csv("PPP.csv")
attach(data1)
x <- data1[[03]]
y <- data1[[04]]
plot(x,y)#to check the data is the data looks OK
data2 <- ppp(x, y, xrange=c(0,464), yrange=c(0,377))#note dimensions of
your image
pcf1 = pcf(data2)
plot(pcf1)

#end



On 15/07/15 20:05, Xiao Zhong wrote:

> Hi everyone,
>
> Thanks for the reply. I have attached a piece of my data here. I just
> wanted to calculate the pair correlation function of these particles
> (g(r)). Since I know different packings would give different g(r).  This
> attached image has spheres roughly form a 2D close packing, which should
> give a g(r) that is different than other types of packings.
>
> I'd appreciate any inputs on this issue.
>
> Thanks,
> Xiao
>
> Xiao Zhong
> Ph.D. Candidate in Chemistry
> Department of Chemistry and Molecular Design Institute
> New York University
>
> On Wed, Jul 15, 2015 at 11:58 AM, Herbie <[hidden email]> wrote:
>
>> Dear Richard,
>>
>> the poster Xiao is not very clear, or at least I don't know what "packed
>> particles" are.
>>
>> If "pair correlation function of a certain 2D area" means pixel-pair
>> correlation in that area, I'm pretty sure that this spatial statistic is
>> identical to the auto-correlation function of the image in this area.
>>
>> Best
>>
>> Herbie
>>
>> ::::::::::::::::::::::::::::::::::::::::::
>> Am 15.07.15 um 17:38 schrieb Richard Mort:
>>
>>   I think in this context Xiao is looking for the spatial statistic. I'm
>>> not aware of an IJ plugin to do this at the moment but it is on my to do
>>> list as its a very useful method. The quickest approach would be to
>>> segment your images and then use analyze particles to get the XY
>>> positions of each object. Save as .csv and then import this data into R
>>> and use the spatstat package:
>>>
>>> http://www.inside-r.org/packages/cran/spatstat/docs/pcf
>>>
>>> Best
>>> R
>>>
>>> On 15/07/15 16:16, Herbie wrote:
>>>
>>>> Xiao,
>>>>
>>>> it's always a good idea to first consult the user guide...
>>>>
>>>> You may have a look at page
>>>>
>>>> <http://rsb.info.nih.gov/ij/docs/guide/146-29.html#toc-Subsection-29.10>
>>>>
>>>> Subsection " 29.10.7 FD Math… "
>>>>
>>>> If you use Image1 = Image2, i.e. the same image, then you get the
>>>> auto-correlation function and I guess this is what you want.
>>>>
>>>> HTH
>>>>
>>>> Herbie
>>>>
>>>> ::::::::::::::::::::::::::::::::::::::::
>>>> Am 15.07.15 um 15:29 schrieb Xiao Zhong:
>>>>
>>>>> Hi everyone,
>>>>>
>>>>> I am Ph.D. candidate and am working on a project concerning colloidal
>>>>> particle packing. I am new to ImageJ and I was wondering if anyone knows
>>>>> how to use it to calculate pair correlation function of a certain 2D
>>>>> area
>>>>> with packed particles.
>>>>>
>>>>> Thanks!
>>>>>
>>>>> Best,
>>>>> Xiao
>>>>>
>>>>> --
>>>>> 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
>>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
--
Dr Richard Mort
Research Fellow
MRC Human Genetics Unit
Institute of Genetics and Molecular Medicine
University of Edinburgh
Western General Hospital
Crewe Road
Edinburgh
EH4 2XU

Tel: 0131 651 8648


The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.


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

pcf.png (71K) Download Attachment
data.png (388K) Download Attachment