Local Background Subtraction

[Joel Sheffield]

Greetings all,

We are interested in quantitating the intensity of spots in dot blot
experiments in which the background is uneven.  Some of the commercial
software packages have instituted a routine that essentially thresholds
each spot, dilates the area by a fixed amount, calculates the per pixel
density in the expanded region, and then subtracts that value from each
pixel in the sample area.

A quick scan through the mailing list, as well as the plugins list, didn't
come up with anything.

Is anyone aware of such a plugin/macro for IJ?

Joel


Joel B. Sheffield, Ph.D
Department of Biology
Temple University
Philadelphia, PA 19122
Voice: 215 204 8839
e-mail: [hidden email]
URL:  *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>*

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

[Ignacio Arganda-Carreras-2]

Dear Joel,

That is similar to applying a White Top Hat filter with a large structuring
element. Have a look at that morphological filter option in MorphoLibJ
<http://imagej.net/MorphoLibJ#Top-hats>.

Cheers!

ignacio

On Sun, Nov 13, 2016 at 6:13 PM, JOEL B. SHEFFIELD <[hidden email]> wrote:

> Greetings all,
>
> We are interested in quantitating the intensity of spots in dot blot
> experiments in which the background is uneven.  Some of the commercial
> software packages have instituted a routine that essentially thresholds
> each spot, dilates the area by a fixed amount, calculates the per pixel
> density in the expanded region, and then subtracts that value from each
> pixel in the sample area.
>
> A quick scan through the mailing list, as well as the plugins list, didn't
> come up with anything.
>
> Is anyone aware of such a plugin/macro for IJ?
>
> Joel
>
>
> Joel B. Sheffield, Ph.D
> Department of Biology
> Temple University
> Philadelphia, PA 19122
> Voice: 215 204 8839
> e-mail: [hidden email]
> URL:  *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>*
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
Ignacio Arganda-Carreras, Ph.D.
Ikerbasque Research Fellow
Departamento de Ciencia de la Computacion e Inteligencia Artificial
Facultad de Informatica, Universidad del Pais Vasco
Paseo de Manuel Lardizabal, 1
20018 Donostia-San Sebastian
Guipuzcoa, Spain

Phone : [phone]
Website: http://sites.google.com/site/iargandacarreras/
<http://biocomp.cnb.csic.es/~iarganda/index_EN.html>

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

[Gabriel Landini]

On Sunday 13 Nov 2016 12:13:43 JOEL B. SHEFFIELD wrote:
> We are interested in quantitating the intensity of spots in dot blot
> experiments in which the background is uneven.  Some of the commercial
> software packages have instituted a routine that essentially thresholds
> each spot, dilates the area by a fixed amount, calculates the per pixel
> density in the expanded region, and then subtracts that value from each
> pixel in the sample area.

Hi Joel,  Are your spots all the same size? (like the dot-blot sample image?).

Regards

Gabriel

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

[Joel Sheffield]

Thanks Ignacio and Gabriel,

I'll try the tophat filter.  Meanwhile, for Gabriel, the spots can vary in
size.  I was also thinking of this as another approach to gel analysis, and
so I was hoping for some flexibility in defining the shape.

Joel



Joel B. Sheffield, Ph.D
Department of Biology
Temple University
Philadelphia, PA 19122
Voice: 215 204 8839
e-mail: [hidden email]
URL:  *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>*

On Mon, Nov 14, 2016 at 5:27 AM, Gabriel Landini <[hidden email]>
wrote:

> On Sunday 13 Nov 2016 12:13:43 JOEL B. SHEFFIELD wrote:
> > We are interested in quantitating the intensity of spots in dot blot
> > experiments in which the background is uneven.  Some of the commercial
> > software packages have instituted a routine that essentially thresholds
> > each spot, dilates the area by a fixed amount, calculates the per pixel
> > density in the expanded region, and then subtracts that value from each
> > pixel in the sample area.
>
> Hi Joel,  Are your spots all the same size? (like the dot-blot sample
> image?).
>
> Regards
>
> Gabriel
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[Ignacio Arganda-Carreras-2]

Hello again, Joel,

If you post an image here or in the ImageJ forum <http://forum.imagej.net/>,
we can advice you on the best method available in ImageJ to solve your
problem.

ignacio

On Mon, Nov 14, 2016 at 2:34 PM, JOEL B. SHEFFIELD <[hidden email]> wrote:

> Thanks Ignacio and Gabriel,
>
> I'll try the tophat filter.  Meanwhile, for Gabriel, the spots can vary in
> size.  I was also thinking of this as another approach to gel analysis, and
> so I was hoping for some flexibility in defining the shape.
>
> Joel
>
>
>
> Joel B. Sheffield, Ph.D
> Department of Biology
> Temple University
> Philadelphia, PA 19122
> Voice: 215 204 8839
> e-mail: [hidden email]
> URL:  *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>*
>
> On Mon, Nov 14, 2016 at 5:27 AM, Gabriel Landini <[hidden email]>
> wrote:
>
> > On Sunday 13 Nov 2016 12:13:43 JOEL B. SHEFFIELD wrote:
> > > We are interested in quantitating the intensity of spots in dot blot
> > > experiments in which the background is uneven.  Some of the commercial
> > > software packages have instituted a routine that essentially thresholds
> > > each spot, dilates the area by a fixed amount, calculates the per pixel
> > > density in the expanded region, and then subtracts that value from each
> > > pixel in the sample area.
> >
> > Hi Joel,  Are your spots all the same size? (like the dot-blot sample
> > image?).
> >
> > Regards
> >
> > Gabriel
> >
> > --
> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
Ignacio Arganda-Carreras, Ph.D.
Ikerbasque Research Fellow
Departamento de Ciencia de la Computacion e Inteligencia Artificial
Facultad de Informatica, Universidad del Pais Vasco
Paseo de Manuel Lardizabal, 1
20018 Donostia-San Sebastian
Guipuzcoa, Spain

Phone : [phone]
Website: http://sites.google.com/site/iargandacarreras/
<http://biocomp.cnb.csic.es/~iarganda/index_EN.html>

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

[Gabriel Landini]

On Monday 14 Nov 2016 08:34:34 JOEL B. SHEFFIELD wrote:
> Meanwhile, for Gabriel, the spots can vary in
> size.  I was also thinking of this as another approach to gel analysis, and
> so I was hoping for some flexibility in defining the shape.

Hi Joel,
If you can segment the spots (eg with a local or regional thresholding
method), then add the segmented image to the original (here I imagine the
masked pixels being 255 and the background 0). The spots now have a value of
255 and the background same as it was before.

Then on this image run a median filter large enough so that the spots just
disappear (because the spots have a value of 255 (an extreme value) they never
appear in the median filter result if the radius is large enough.

If you use that result as "background" you can then divide the original by
this computed background image using the Calculator Plus plugin.
Use the original as i1 and divide by "background" (i2) and for the multiplier
use 255.
That should remove the uneven background by the transmittance method with the
medial filtered image as an estimation of the background.

You'll need to download the latest Morphology zip from my site or the Fiji
Morphology update site to run this, and mind the line breaks:

// -----------8<-------------
run("Dot Blot (7K)");
run("Gaussian Blur...", "sigma=1");

run("Threshold Regional Gradient", "circularity=0.60 minimum=400 maximum=3600
fill_phase fill_detected method=Fast");

selectWindow("Result");
rename("bin");
imageCalculator("Add create", "Dot_Blot.jpg","bin");
selectWindow("Result of Dot_Blot.jpg");
run("Median...", "radius=21");

run("Calculator Plus", "i1=Dot_Blot.jpg i2=[Result of Dot_Blot.jpg]
operation=[Divide: i2 = (i1/i2) x k1 + k2] k1=255 k2=0 create");

selectWindow("bin");
run("NumberParticles8 ");
selectWindow("bin");

run("Particles8 ", "white show=Particles minimum=0 maximum=9999999 display
overwrite redirect=Result");
// -----------8<-------------

Note that the regions get numbered from 0 and the table from 1 and that the
initial segmentation expects round regions of a given range of sizes. Maybe
you can adapt the above to your images.
Hope it helps

Gabriel

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