Re: IMAGEJ Digest - 3 Oct 2013 to 4 Oct 2013 (#2013-288)

[chaitanya athale]

Dear all,
we were using Analyze->Analyze particles and find something strange. When
we use the size-cutoff parameter (upper-lower) threshold to eliminate
particles of a certain area, it works fine. But using the circularity
parameter as a cutoff doesn't work as well. I have attached the example
image I was testing for this. The small circle (area cutoff) gets
eliminated if the lower bound of area is set to 300. But circularity upper
bound set to 0.7 does not eliminate the 2 big circles. Anybody with ideas
do please drop me a line.

Just to clarify, we are running ImageJ 1.45r on both windows8 and Mac OSX.


Regards,
Chaitanya Athale.
IISER Pune, India

Attached: 8 bit TIF image of white objects against black background. They
were threholded using the Image->Set Threshold command.


On Sat, Oct 5, 2013 at 9:30 AM, IMAGEJ automatic digest system <
[hidden email]> wrote:

> There are 6 messages totaling 468 lines in this issue.
>
> Topics of the day:
>
>   1. Pls help me with segmentation issue to count cell colonies in 3D
> agars.
>      (2)
>   2. Segmentation issue
>   3. FRAP analysis help please (2)
>   4. Memory issue
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> ----------------------------------------------------------------------
>
> Date:    Fri, 4 Oct 2013 10:32:42 +0200
> From:    Peter Haub <[hidden email]>
> Subject: Re: Pls help me with segmentation issue to count cell colonies in
> 3D agars.
>
> Hi Alan,
>
> you can find various infos about shading correction (or flat field
> correction) in the net. Search for"ImageJ shading correction".
> Examples are:
> http://rsbweb.nih.gov/ij/plugins/shading-corrector.html
>
> http://imagejdocu.tudor.lu/doku.php?id=howto:working:how_to_correct_background_illumination_in_brightfield_microscopy
>
> The background in your images are so inhomogeneous that thresholding
> will lead to problems.
> The problem will be that you do not have a background image since the
> images are captured month ago - as you wrote.
>
> The images look like phase contrast. I would try to eliminate the ring
> artefact around the objects. But this is an academic discussion if there
> is no chance to recapture the samples.
>
> Just one more hint: Saving scientific images in jpg-format is not
> optimal. You will lose information and introduce jpg-artefact which are
> clearly visible in your images and which are sub-optimal for every
> analysis process.
>
> Regards,
> Peter
>
> On 02.10.2013 21:44, Alan Thames wrote:
> > Hi Peter,
> >                 I used Nikon Eclipse TS100 microscope with Inifinity 2
> > capturing tool from Micron-Optics. I took these pictures months ago and
> > saved as jpeg formats.  Where can i find how to do a flat field
> correction?
> >
> > Thank you.
> >
> > Alan
> >
> >
> > On Mon, Sep 30, 2013 at 2:12 PM, Peter Haub <[hidden email]> wrote:
> >
> >> Hi Alan,
> >>
> >> you can try to use a minimum filter (Process > Filters > Minimum ..).
> >> Maybe this can help to separate (and count) the colonies.
> >>
> >> In any case you should do a flat field correction.
> >>
> >> And maybe there is a chance to improve the imaging. What's your optical
> >> system?
> >>
> >> Regards,
> >> Peter
> >>
> >>
> >> On 30.09.2013 18:07, Alan Thames wrote:
> >>
> >>> Dear ImageJ users,
> >>>                                I have an issue with counting my cell
> >>> colonies
> >>> in 3D agars. I have attached 8 bit images, in which you can see control
> >>> image
> >>> has huge and multiple colonies, while the other image  treated with a
> >>> drug,
> >>>    50 uM concentation has  multiple tiny
> >>> colonies. When i try to count the colonies number with imageJ, i found
> >>>    the count
> >>> " number" is even  more increased  in treated images, which i do not
> want
> >>> to see,obviously, though the size of the colonies are counted to be
> >>> decreased with increasing drug concentration ( pls see below results;
> >>> settings - Image>Adjust>Threshold>Auto, Analyze>Analyze Particles,
> Size: 0
> >>> to infinity, circularity 0-1 and Exclude on Edges)  . As one suggested,
> >>> the
> >>> colonies are all in different planes (ie the agar is too tick to b
> >>> enabling
> >>> focusing
> >>> in all colonies at once). That will prevent segmenting them right as
> >>> there are
> >>> too many halo (out of focus) artifacts. Can any one help me with this
> >>> please?
> >>>
> >>>
> >>>                           Count               total Area
> Average
> >>> Size
> >>>          % area
> >>> Control:             4164 398785.000000 95.769693 12.482002
> >>> 50 uM :             9691   194228.000000 20.042101 6.079352
> >>>
> >>>
> >>> I appreciate it.
> >>>
> >>> Regards,
> >>> Alan
> >>>
> >>> --
> >>> ImageJ mailing list: http://imagej.nih.gov/ij/list.**html<
> http://imagej.nih.gov/ij/list.html>
> >>>
> >> --
> >> ImageJ mailing list: http://imagej.nih.gov/ij/list.**html<
> 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
>
> ------------------------------
>
> Date:    Fri, 4 Oct 2013 12:22:20 +0200
> From:    Peter Haub <[hidden email]>
> Subject: Re: Segmentation issue
>
> Dear Alan,
>
> attached you find a simple macro which creates a background images,
> performs a 'flat field' correction, applies a threshold and count the
> particles with the Particle Analyzer.
>
> The results I got with this macro are (size=300-Infinity):
>
> Slice                     Count         Total Area        Average Size
>     %Area       IntDen
> control                    180        498631.000       2770.172
> 15.607    706393.917
> 10 micromolar        147        259839.000       1767.612 8.133
> 450741.122
> 50 micromolar        121        82643.000          683.000     2.587
>      174165
>
> Hope this will help you to find a way into your work.
>
> Regards,
> Peter
>
> On 02.10.2013 23:42, Alan Thames wrote:
> >   Hi Peter,
> >                 Sorry that i had to write a direct mail to you as i
> cannot
> > attach many images to imageJ mailing list. As you have suggested, i tried
> > to use filter as Process > filter>minimum - Radius 2.0 pixels for each 8
> > bit image - control and 10 micromolar concentration . Then i followed my
> > previous step - choosing IMAGE>adjust > threshold. I do not change
> anything
> > under threshold setting. After that i go to Analyze> analyze particles >
> > size 0 - infinity, circularity 0.01-1.0/ 0.5 - 1.0 , then chose -exclude
> on
> > edges. the summary of the result
> >
> > circularity 0.01-1.0
> >                               count   total area           average size
> >   percent area
> > control                    681 448209.000000 658.162996 14.028978
> > 10 uM                    1089 293413.000000 269.433425 9.183850
> >
> > circularity 0.5 - 1.0
> >
> > control                     614 230438.000000 375.306189 7.212728
> >     10 uM                   978 174331.000000 178.252556 5.456574
> >
> >
> > I have attached three pictures , control, 10 uM and 50 uM drug
> > concentrations.
> >
> > I would also like to learn how to do flat field correction.
> > Will you please take a look at what i have done above? i have been
> having a
> > hard time figuring out how to do this as i want to see decrease in both
> >   count and size with increasing drug concentrations. This is part of my
> > masters thesis project. I cannot one any one near me.
> >
> > Sincerely,
> > Alan
> >
>
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> ------------------------------
>
> Date:    Fri, 4 Oct 2013 03:33:06 -0700
> From:    RUNinspired <[hidden email]>
> Subject: FRAP analysis help please
>
> My biggest pondering related to FRAP deals with the following scenarios. I
> am
> dealing with FRAP analysis on moving axons in one instance or just moving
> puncta within them. Both shapes are very tricky compared to a classical
> textbook FRAP of a small rectangle within a cell. My axons are dynamic and
> in the process of recovering they change shape, length etc. The puncta as
> well goes up and down the axon.
>
>
> I guess my first question is about the more simple scenario of the moving
> puncta. Before bleaching the puncta has an area X and after the recovery
> the
> area never reaches the size X and it always remains smaller. *What area do
> I
> use to collect a mean value? I am afraid that by selecting just the area of
> the puncta I am artificially increasing the mean value of recovery. * After
> all a big and a small area can have very close mean values despite the fact
> that the smaller area has obviously lost part of its fluorescence.* Do I
> use
> a selection from the first pre-bleach frame and keep using this selection,
> or do I manually go and track each frame since the puncta grows slowly over
> time. Is there any way to adjust for area-size?* What about using total sum
> of pixels? I am attaching a little image. (
> http://i39.tinypic.com/2ir2a1c.jpg <http://i39.tinypic.com/2ir2a1c.jpg>  )
> In this example the puncta did not move much but very often they would
> traverse almost the full field of view. When I took this project I thought
> tracking puncta must be the simplest thing to do since they are well
> defined
> but I have not had much luck in automating FRAP analysis on this recovering
> moving puncta, Especially since tracking fails if there are higher number
> puncta and some are bleached and some are not.
>
> *-Summary: FRAP on puncta which are moving a lot and also growing over time
> after the recovery. What area do I use? If I use area in pre-FRAP, the
> recovering puncta never come back to the same big size. If I have different
> area for each frame (smaller area) the mean value is basically increased a
> lot since removing the empty volume around. 1 px of 1000 units has the same
> mean as 100px of 100 000 total units.*
>
> My second most important question is how should I go about measuring FRAP
> on
> the dynamic axons. I am attaching a sample image. The axons shorten,
> elongate, sometimes quite significantly. My previous question about the
> shrinking area is still valid. Very often the recovered axon area of signal
> is slightly thinner than the pre-bleach.  Is my best bet to basically
> manually trace out each individual frame and measure mean. Would it be
> possible to have any type of automation on such a problem?
> http://i41.tinypic.com/219vr5z.jpg <http://i41.tinypic.com/219vr5z.jpg>
>
> *-Summary: Second question is very similar to the first as in dealing with
> a
> shape in FRAP that is changing size, shape, etc*
>
> THANK YOU!!!!
>
>
>
> --
> View this message in context:
> http://imagej.1557.x6.nabble.com/FRAP-analysis-help-please-tp5005040.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> ------------------------------
>
> Date:    Fri, 4 Oct 2013 14:37:16 +0200
> From:    Johannes Schindelin <[hidden email]>
> Subject: Re: Memory issue
>
> Hi Andrew,
>
> On Thu, 3 Oct 2013, Andrew Riching wrote:
>
> > Alright, I've tested out a few things now. While I was at work, I tried
> > opening files as virtual stacks on Windows XP, Windows 7, and MacOS
> > 10.8.3.  XP and 7 were significantly slower (it probably took between 3
> > to 5 minutes to open the file on either OS). MacOS 10.8.3 took maybe 30
> > seconds to open the file as a virtual stack.
>
> That sounds consistent with a virus scanner (required on Windows, but
> maybe not so much for .tif files).
>
> Ciao,
> Johannes
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> ------------------------------
>
> Date:    Fri, 4 Oct 2013 14:50:23 +0200
> From:    Johannes Schindelin <[hidden email]>
> Subject: Re: Pls help me with segmentation issue to count cell colonies in
> 3D agars.
>
> Hi Peter,
>
> On Fri, 4 Oct 2013, Peter Haub wrote:
>
> > Just one more hint: Saving scientific images in jpg-format is not
> > optimal. You will lose information and introduce jpg-artefact which are
> > clearly visible in your images and which are sub-optimal for every
> > analysis process.
>
> Often, those artifacts are not so clearly visible (JPEG format was
> designed to fool our eyes, after all...) unless after processing -- and
> then often only to the trained eye.
>
> But artifacts they are, therefore JPEG is a poor format to store
> scientific images in. It is appropriate only in a very narrow set of
> bandwidth-constrained visualization tasks, and then only for transmitting,
> never for storing nor analyzing.
>
> Ciao,
> Johannes
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> ------------------------------
>
> Date:    Fri, 4 Oct 2013 19:02:00 +0000
> From:    "Unruh, Jay" <[hidden email]>
> Subject: Re: FRAP analysis help please
>
> From the images you uploaded it looks like the "area" changes after
> bleaching are simply due to saturation before bleaching and lack thereof
> afterwards.  All objects will look bigger in a saturated image.  Are you
> trying to measure exchange of the red label to and from this structure?  In
> that case, the best option is to measure the raw integrated density from
> the object (in a background subtracted sum projected z stack ideally) as a
> function of time during the recovery.  In this analysis, area doesn't
> matter because the recovery is measured in terms of total label, not the
> area that it occupies.  If you care about the fraction of mobile red label,
> make sure you account for the relative about of red label you bleached
> initially.  Also avoid saturation.
>
> Judging from the size of your image, this object is approaching the
> resolution limit anyway in which case you really have no idea what the
> actual area is.
>
> Alternatively you might want to measure the motion of the red label within
> the structure.  This is more difficult, especially for this small size
> structure.  In that case, you could bleach half of the structure (hopefully
> your scope is accurate enough to do this) and watch the decay of
> fluorescence from the unbleached part and the fluorescence increase in the
> bleached part.  In this case the area will change over time but it doesn't
> necessarily need to be measured in the analysis.
>
> In the case of your whole-axon recovery experiments, I would recommend
> bleaching a much smaller portion of the axon initially.  That will decrease
> the recovery time so that you avoid morphological change.
>
> Jay
>
> -----Original Message-----
> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of
> RUNinspired
> Sent: Friday, October 04, 2013 5:33 AM
> To: [hidden email]
> Subject: FRAP analysis help please
>
> My biggest pondering related to FRAP deals with the following scenarios. I
> am dealing with FRAP analysis on moving axons in one instance or just
> moving puncta within them. Both shapes are very tricky compared to a
> classical textbook FRAP of a small rectangle within a cell. My axons are
> dynamic and in the process of recovering they change shape, length etc. The
> puncta as well goes up and down the axon.
>
>
> I guess my first question is about the more simple scenario of the moving
> puncta. Before bleaching the puncta has an area X and after the recovery
> the area never reaches the size X and it always remains smaller. *What area
> do I use to collect a mean value? I am afraid that by selecting just the
> area of the puncta I am artificially increasing the mean value of recovery.
> * After all a big and a small area can have very close mean values despite
> the fact that the smaller area has obviously lost part of its
> fluorescence.* Do I use a selection from the first pre-bleach frame and
> keep using this selection, or do I manually go and track each frame since
> the puncta grows slowly over time. Is there any way to adjust for
> area-size?* What about using total sum of pixels? I am attaching a little
> image. ( http://i39.tinypic.com/2ir2a1c.jpg <
> http://i39.tinypic.com/2ir2a1c.jpg>  ) In this example the puncta did not
> move much but very often they would traverse almost the full field of view.
> When I took this project I thought tracking puncta must be the simplest
> thing to do since they are well defined but I have not had much luck in
> automating FRAP analysis on this recovering moving puncta, Especially since
> tracking fails if there are higher number puncta and some are bleached and
> some are not.
>
> *-Summary: FRAP on puncta which are moving a lot and also growing over
> time after the recovery. What area do I use? If I use area in pre-FRAP, the
> recovering puncta never come back to the same big size. If I have different
> area for each frame (smaller area) the mean value is basically increased a
> lot since removing the empty volume around. 1 px of 1000 units has the same
> mean as 100px of 100 000 total units.*
>
> My second most important question is how should I go about measuring FRAP
> on the dynamic axons. I am attaching a sample image. The axons shorten,
> elongate, sometimes quite significantly. My previous question about the
> shrinking area is still valid. Very often the recovered axon area of signal
> is slightly thinner than the pre-bleach.  Is my best bet to basically
> manually trace out each individual frame and measure mean. Would it be
> possible to have any type of automation on such a problem?
> http://i41.tinypic.com/219vr5z.jpg <http://i41.tinypic.com/219vr5z.jpg>
>
> *-Summary: Second question is very similar to the first as in dealing with
> a shape in FRAP that is changing size, shape, etc*
>
> THANK YOU!!!!
>
>
>
> --
> View this message in context:
> http://imagej.1557.x6.nabble.com/FRAP-analysis-help-please-tp5005040.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
>
> ------------------------------
>
> End of IMAGEJ Digest - 3 Oct 2013 to 4 Oct 2013 (#2013-288)
> ***********************************************************
>

--
-------------------------
Chaitanya Athale,
Pune, India
18° 31' N, 73° 55' E, 560m.
-------------------------

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