Measurement of concentric circles

[Sara]

Hello,

I'm trying to measure the area within a number of concentric circles (the circles represent daily fungal growth in a petri dish, which get larger each day, radiating out from a central 5mm square).

I understand how to convert to RGB (red) and how to get ImageJ to calc the total red area, but not sure how to pick up the black circles I've hand drawn on the plate present in the image.  Do I need to manually draw around these with the polygon tool?  I'm hoping that there is a (?)plugin that can recognise them in the image, and that I can assign a ?label to each circle before calculating area.

Im very new to this program and not terribly tech-savvy, so any help is greatly appreciated.

Thanks,
Sara

[Kischell, Eric R.]

Sara,
Could you please send some sample images?
Have you looked at the Hough Transform?
Please search for "circle" or "Hough" on http://rsbweb.nih.gov/ij/plugins/ .
There seems to be a few interesting plugins.
Regards,
e.-

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Sara
Sent: Thursday, June 13, 2013 6:49 PM
To: [hidden email]
Subject: Measurement of concentric circles

Hello,

I'm trying to measure the area within a number of concentric circles (the circles represent daily fungal growth in a petri dish, which get larger each day, radiating out from a central 5mm square).

I understand how to convert to RGB (red) and how to get ImageJ to calc the total red area, but not sure how to pick up the black circles I've hand drawn on the plate present in the image.  Do I need to manually draw around these with the polygon tool?  I'm hoping that there is a (?)plugin that can recognise them in the image, and that I can assign a ?label to each circle before calculating area.

Im very new to this program and not terribly tech-savvy, so any help is greatly appreciated.

Thanks,
Sara



--
View this message in context: http://imagej.1557.x6.nabble.com/Measurement-of-concentric-circles-tp5003435.html
Sent from the ImageJ mailing list archive at Nabble.com.

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

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

[Sara]

Thanks for the advice - I'm not sure that Hough's is what I'm looking for?

Attached is an example pic.  I can scan in at higher res if required.

I'm hoping ImageJ can identify each black line and then calculate the area within each shape.  I realise that the thickness of the black line may be an issue, so consistent inclusion of the area the line occupies in either of the adjoining donut shapes is fine.

Really appreciate the help :)

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

[Graeme Ball-2]

Hi Sara,

Which black lines? only the central 6 rings? or the less well-defined ones
at the edge too?

Also, is this something you want to do more than say tens of times?


If you only want the central rings, and don't need to repeat it a huge
number of times (i.e. some manual input is bearable), the following steps
should work (recorded with macro recorder):-

run("Invert");  // Edit -> Invert
run("Bandpass Filter...", "filter_large=12 filter_small=4 suppress=None
tolerance=5 autoscale saturate");  // Process -> FFT -> Bandpass
setAutoThreshold("Triangle dark");  // Image -> Adjust -> Threshold  (tick
dark background, use triangle method)
//run("Threshold...");
setThreshold(111, 255);
setOption("BlackBackground", true);
run("Convert to Mask");  // apply the threshold to create a binary mask
image
run("Skeletonize");

Now open the ROI manager (Analyze->Tools->ROI manager), and use the wand
tool to select the central ring. Add to the ROI manager with "t". Use "m"
key to measure area.

You can find the areas between the rings by selecting pairs of rings (hold
down shift and select two ROIs in ROI manager), and using the "More>> XOR"
operation in the ROI manager to create a ROI for the region between the
rings. Again, use "t" to add the new ROI to the manager, "m" to measure it.

Of course all this could be automated, and there are more sophisticated
things you could do, but it should give you a reasonable first answer.

Regards,

Graeme


On Sat, Jun 15, 2013 at 2:12 AM, Sara <[hidden email]> wrote:

> Thanks for the advice - I'm not sure that Hough's is what I'm looking for?
>
> Attached is an example pic.  I can scan in at higher res if required.
>
> I'm hoping ImageJ can identify each black line and then calculate the area
> within each shape.  I realise that the thickness of the black line may be
> an issue, so consistent inclusion of the area the line occupies in either
> of the adjoining donut shapes is fine.
>
> Really appreciate the help :)
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[Sara]

Dear Graeme,

Thankyou!  This is absolutely what I was after.  I really appreciate your help on this - its going to save me from so much manual (ruler-based) measurement.  The small amount of manual input is fine as I need to be able to calculate the (growth) difference between each circle which I can do in excel (plus a bunch of other stats).

Again, thank you so much - I'd like to acknowledge you in my thesis for technical assistance, if that is ok by you?

Best regards,
Sara

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

[Gabriel Landini]

On Saturday 15 Jun 2013 02:12:53 Sara wrote:
> Thanks for the advice - I'm not sure that Hough's is what I'm looking for?
>
> Attached is an example pic.  I can scan in at higher res if required.
> I'm hoping ImageJ can identify each black line and then calculate the area
> within each shape.  I realise that the thickness of the black line may be
> an issue, so consistent inclusion of the area the line occupies in either
> of the adjoining donut shapes is fine.

Sorry to give unsolicited advice and maybe I do not understand the problem
well, but the approach does not seem ideal to me.

If this is to measure some migration from the rectangle in the centre, what I
would do instead is first identify the square-ish area in the dish (by some
segmentation method), then compute the distance transform of the background
and use that for defining the distance intervals from the rectangle. You could
use a threshold of the distance map to define a particular distance. The shape
of the intervals will also will transition correctly from a square to near
circles when they become larger.
Hope it is useful and sorry if I misunderstood what you are trying to achieve.
Regards

Gabriel

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

[Sara]

Hi Gabriel,


Thank you for your time , I do appreciate your offer of help.  Someone else was able to help me out - I'm not terribly tech-savvy either which probably made this more difficult.

Best regards,
Sara



----- Original Message -----
From: "Gabriel Landini [via ImageJ]" <[hidden email]>
Date: Saturday, June 15, 2013 19:39
Subject: Re: Measurement of concentric circles
To: Sara <[hidden email]>

> On Saturday 15 Jun 2013 02:12:53 Sara wrote:
> > Thanks for the advice - I'm not sure that Hough's is what I'm looking for?
> >
> > Attached is an example pic.  I can scan in at higher res if required.
> > I'm hoping ImageJ can identify each black line and then calculate the area
> > within each shape.  I realise that the thickness of the black line may be
> > an issue, so consistent inclusion of the area the line occupies in either
> > of the adjoining donut shapes is fine.
>
> Sorry to give unsolicited advice and maybe I do not understand the problem
> well, but the approach does not seem ideal to me.
>
> If this is to measure some migration from the rectangle in the centre, what I
> would do instead is first identify the square-ish area in the dish (by some
> segmentation method), then compute the distance transform of the background
> and use that for defining the distance intervals from the rectangle. You could
> use a threshold of the distance map to define a particular distance. The shape
> of the intervals will also will transition correctly from a square to near
> circles when they become larger.
> Hope it is useful and sorry if I misunderstood what you are trying to achieve. 



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

---

If you reply to this email, your message will be added to the discussion below:

> http://imagej.1557.x6.nabble.com/Measurement-of-concentric-circles-tp5003435p5003475.html

>

To unsubscribe from Measurement of concentric circles, click here.
> NAML