Circularity error?

[Adam Hughes]

Hi,

I made an image of increasing circles to see how the circularity changed as
a function of circle radius, where I used another library to draw the
circles.

I discovered that if I compute the circularity using the formula  4 pi  * (
area / perim**2), the result is diferent than if I plot the output of
circularity.  All of the values of area, perimeter and circularity were
taken directly from imageJ analyze particles.

It looks like the circularty has some sort of correction in place?

I've attached the scatter pot showing these results and the test data used
for the particles.

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

[Cammer, Michael]

Hi.

I took your image of circles, thresholded it, and ran particle analyzer on it.  The resulting calculated circularity looks different than your results (attached).  Did you draw a circular ROI and measure it as opposed to the raster based circles?  Regardsless, the Roundness was 1 for all the circles.

Regards,

Michael

===========================================================================
Michael Cammer, Microscopy Core & Dustin Lab , Skirball Institute, NYU Langone Medical Center
Cell:  914-309-3270   Lab: 212-263-3208
http://ocs.med.nyu.edu/microscopy & http://www.med.nyu.edu/skirball-lab/dustinlab/


-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Adam Hughes
Sent: Wednesday, February 19, 2014 1:47 PM
To: [hidden email]
Subject: Circularity error?

Hi,

I made an image of increasing circles to see how the circularity changed as a function of circle radius, where I used another library to draw the circles.

I discovered that if I compute the circularity using the formula  4 pi  * ( area / perim**2), the result is diferent than if I plot the output of circularity.  All of the values of area, perimeter and circularity were taken directly from imageJ analyze particles.

It looks like the circularty has some sort of correction in place?

I've attached the scatter pot showing these results and the test data used for the particles.

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

------------------------------------------------------------
This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain information that is proprietary, confidential, and exempt from disclosure under applicable law. Any unauthorized review, use, disclosure, or distribution is prohibited. If you have received this email in error please notify the sender by return email and delete the original message. Please note, the recipient should check this email and any attachments for the presence of viruses. The organization accepts no liability for any damage caused by any virus transmitted by this email.
=================================

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

[Cammer, Michael]

Also, for one of the circles...
                                        Area (pixels) Perim. Circ. Round
magic wand 2809 196.17 0.92 1
magic wand followed by Fit Spline 2756 187.3 0.99 0.99
magic wand followed by Fit Circle 2828 188.5 1 1
magic wand followed by Fit Ellipse 2809 187.82 1 1

===========================================================================
Michael Cammer, Microscopy Core & Dustin Lab , Skirball Institute, NYU Langone Medical Center
Cell:  914-309-3270   Lab: 212-263-3208
http://ocs.med.nyu.edu/microscopy & http://www.med.nyu.edu/skirball-lab/dustinlab/


-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Cammer, Michael
Sent: Wednesday, February 19, 2014 2:28 PM
To: [hidden email]
Subject: Re: Circularity error?

Hi.

I took your image of circles, thresholded it, and ran particle analyzer on it.  The resulting calculated circularity looks different than your results (attached).  Did you draw a circular ROI and measure it as opposed to the raster based circles?  Regardsless, the Roundness was 1 for all the circles.

Regards,

Michael

===========================================================================
Michael Cammer, Microscopy Core & Dustin Lab , Skirball Institute, NYU Langone Medical Center
Cell:  914-309-3270   Lab: 212-263-3208
http://ocs.med.nyu.edu/microscopy & http://www.med.nyu.edu/skirball-lab/dustinlab/


-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Adam Hughes
Sent: Wednesday, February 19, 2014 1:47 PM
To: [hidden email]
Subject: Circularity error?

Hi,

I made an image of increasing circles to see how the circularity changed as a function of circle radius, where I used another library to draw the circles.

I discovered that if I compute the circularity using the formula  4 pi  * ( area / perim**2), the result is diferent than if I plot the output of circularity.  All of the values of area, perimeter and circularity were taken directly from imageJ analyze particles.

It looks like the circularty has some sort of correction in place?

I've attached the scatter pot showing these results and the test data used for the particles.

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

------------------------------------------------------------
This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain information that is proprietary, confidential, and exempt from disclosure under applicable law. Any unauthorized review, use, disclosure, or distribution is prohibited. If you have received this email in error please notify the sender by return email and delete the original message. Please note, the recipient should check this email and any attachments for the presence of viruses. The organization accepts no liability for any damage caused by any virus transmitted by this email.
=================================

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

------------------------------------------------------------
This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain information that is proprietary, confidential, and exempt from disclosure under applicable law. Any unauthorized review, use, disclosure, or distribution is prohibited. If you have received this email in error please notify the sender by return email and delete the original message. Please note, the recipient should check this email and any attachments for the presence of viruses. The organization accepts no liability for any damage caused by any virus transmitted by this email.
=================================

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

[Gabriel Landini]

On Wednesday 19 Feb 2014 19:28:19 Cammer, Michael wrote:
> I took your image of circles, thresholded it, and ran particle analyzer on
> it.  The resulting calculated circularity looks different than your results
> (attached).  Did you draw a circular ROI and measure it as opposed to the
> raster based circles?  Regardsless, the Roundness was 1 for all the
> circles.

You can't get circularity of 1 on a square lattice, because you can only
approximate a circle. The error varies with the size of the region (and the
encoding methods used to compute for perimeter and area).
I think the particle analyzer in IJ does (or used to do) some correction for
small size regions. That might be desirable when filtering regions by
circularity.

If you want to get the non-corrected values, the Particles8 plugin computes
those.

Regards

Gabriel

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

[Adam Hughes]

Michael,

Thanks.  I also just thresholded and used the rasterised circles.  I've
since tried again in FIJI to ensure the results were the same, and they
are.  As Gabriel pointed out, it's unusual that you'd get a circularity of
1 for you smallest particles, no?  The small particles are rectangles and
should have the least circularity in principle, although the encoding
method, as Gabrield said, can deviate from the geometric fit of that
particluar object.

Gabriel, thanks for pointing out that you think a correction may be in
place.  If anyone knows more about this correction, I'd be interested in
learning a bit about it.

As for the non-corrected values, I would like to try the Particles8 plugin
but can't find a link after a quick search.  Will search futher later;
however, since imj outputs the perimiter and area already, wouldn't the
circularity that I've computed from the equation 4pi * A / P**2  be the
non-corrected value?  This is shown in my plot as the blue points.


On Wed, Feb 19, 2014 at 2:56 PM, Gabriel Landini <[hidden email]>wrote:

> On Wednesday 19 Feb 2014 19:28:19 Cammer, Michael wrote:
> > I took your image of circles, thresholded it, and ran particle analyzer
> on
> > it.  The resulting calculated circularity looks different than your
> results
> > (attached).  Did you draw a circular ROI and measure it as opposed to the
> > raster based circles?  Regardsless, the Roundness was 1 for all the
> > circles.
>
> You can't get circularity of 1 on a square lattice, because you can only
> approximate a circle. The error varies with the size of the region (and the
> encoding methods used to compute for perimeter and area).
> I think the particle analyzer in IJ does (or used to do) some correction
> for
> small size regions. That might be desirable when filtering regions by
> circularity.
>
> If you want to get the non-corrected values, the Particles8 plugin computes
> those.
>
> Regards
>
> Gabriel
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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