Background correction

[anusuya pal]

Dear all,

I have image which has the background issue due to the reflected light from
the sample. To understand whether the mean intensity varies actually due to
sample configuration, I need to first correct the background. Ideally the
mean intensity should be zero as the image is 8 bit with dark background.
But it is not the case for my sample.

I know there are various techniques which correct the background issue in
stacks. I have an idea about only one technique, called BaSiC. Is there any
other way to correct the background in stack?

My current image is not in stack, it is a single image. May I know whether
there is any tool for background correction for single image?

Any suggestions?

Thanks
Anu

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

[Jeremy Adler-2]

My immediate thought would be to add a physical filter to the microscope to reject reflected light.

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of anusuya pal
Sent: 30 September 2017 16:11
To: [hidden email]
Subject: Background correction

Dear all,

I have image which has the background issue due to the reflected light from the sample. To understand whether the mean intensity varies actually due to sample configuration, I need to first correct the background. Ideally the mean intensity should be zero as the image is 8 bit with dark background.
But it is not the case for my sample.

I know there are various techniques which correct the background issue in stacks. I have an idea about only one technique, called BaSiC. Is there any other way to correct the background in stack?

My current image is not in stack, it is a single image. May I know whether there is any tool for background correction for single image?

Any suggestions?

Thanks
Anu

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

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

[Bill Christens-Barry-2]

Anu,

You mention wanting to reject reflected light; am I right in thinking that the light you do want to capture in the image is due to fluorescence? If so, two approaches come to mind:

It might be possible to use linear polarizers. Reflected light often retains the same linear polarization as that of the illumination, while fluorescence emission is often highly depolarized. If you place a linear polarizer in front of the light source and place another in front of the camera, with its polarization axis perpendicular to that of the illumination polarizer, the reflected light will be greatly attenuated while the fluorescence emission will be decreased to a lesser extent (~ 2x for full depolarization). This is especially true if the reflected light has been reflected from the surface alone; translucent materials will allow more penetration and consequent depolarization due to internal scattering, which reduces the effectiveness of this approach. Try rotating the illumination polarizer and the camera polarizer about their polarization axes together, i.e. by maintaining the 90º difference between the orientations of their polarization axes as you rotate both. There will likely be a best orientation for the pair of polarizers.

Alternatively, you might employ a long pass filter in front of the camera so that only the longer emissions are passed while the shorter illumination wavelengths are blocked. This presumes that you are illuminating with a range of wavelengths shorter than those you wish to capture, otherwise some of the illumination may be passed by the filter.

Hope this helps,

Bill Christens-Barry

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

[anusuya pal]

Hello

Here is the sample image (raw file) attached. I am using a polarizing
microscope. And I think for my other samples, BaSiC is working good. But
for particularly this type of image, I am not getting the background
intensity and sample intensity differently. The range is more or less same.
So, if I know any plugins that will solve this correction for the current
image, It will be helpful.

Thanks
Anu

On Sat, Sep 30, 2017 at 12:17 PM, Bill Christens-Barry <[hidden email]>
wrote:

> Anu,
>
> You mention wanting to reject reflected light; am I right in thinking that
> the light you do want to capture in the image is due to fluorescence? If
> so, two approaches come to mind:
>
> It might be possible to use linear polarizers. Reflected light often
> retains the same linear polarization as that of the illumination, while
> fluorescence emission is often highly depolarized. If you place a linear
> polarizer in front of the light source and place another in front of the
> camera, with its polarization axis perpendicular to that of the
> illumination polarizer, the reflected light will be greatly attenuated
> while the fluorescence emission will be decreased to a lesser extent (~ 2x
> for full depolarization). This is especially true if the reflected light
> has been reflected from the surface alone; translucent materials will allow
> more penetration and consequent depolarization due to internal scattering,
> which reduces the effectiveness of this approach. Try rotating the
> illumination polarizer and the camera polarizer about their polarization
> axes together, i.e. by maintaining the 90º difference between the
> orientations of their polarization axes as you rotate both. There will
> likely be a best orientation for the pair of polarizers.
>
> Alternatively, you might employ a long pass filter in front of the camera
> so that only the longer emissions are passed while the shorter illumination
> wavelengths are blocked. This presumes that you are illuminating with a
> range of wavelengths shorter than those you wish to capture, otherwise some
> of the illumination may be passed by the filter.
>
> Hope this helps,
>
> Bill Christens-Barry
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[Herbie]

Good day Anu,

thanks for providing a sample image.

However, I'm still not sure what's wrong with this image with respect to
your further analyses.

1. There is a global background (illumination) gradient in the sample image.

2. There is a light border or hem, especially on the right half, that
follows the object contour.

re 1. As others have noted already: Use better image acquisition
(illumination). Try highly diffuse illumination if possible.
(You may also try "Subtract Background..." but be aware that it affects
your object and its analyses as well.)

re 2. What causes this surprisingly pronounced border?

How big is the object or what magnification did you use?"

Regards

Herbie

:::::::::::::::::::::::::::::::::::::::::
Am 30.09.17 um 20:20 schrieb anusuya pal:

> Hello
>
> Here is the sample image (raw file) attached. I am using a polarizing
> microscope. And I think for my other samples, BaSiC is working good. But
> for particularly this type of image, I am not getting the background
> intensity and sample intensity differently. The range is more or less same.
> So, if I know any plugins that will solve this correction for the current
> image, It will be helpful.
>
> Thanks
> Anu
>
> On Sat, Sep 30, 2017 at 12:17 PM, Bill Christens-Barry <[hidden email]>
> wrote:
>
>> Anu,
>>
>> You mention wanting to reject reflected light; am I right in thinking that
>> the light you do want to capture in the image is due to fluorescence? If
>> so, two approaches come to mind:
>>
>> It might be possible to use linear polarizers. Reflected light often
>> retains the same linear polarization as that of the illumination, while
>> fluorescence emission is often highly depolarized. If you place a linear
>> polarizer in front of the light source and place another in front of the
>> camera, with its polarization axis perpendicular to that of the
>> illumination polarizer, the reflected light will be greatly attenuated
>> while the fluorescence emission will be decreased to a lesser extent (~ 2x
>> for full depolarization). This is especially true if the reflected light
>> has been reflected from the surface alone; translucent materials will allow
>> more penetration and consequent depolarization due to internal scattering,
>> which reduces the effectiveness of this approach. Try rotating the
>> illumination polarizer and the camera polarizer about their polarization
>> axes together, i.e. by maintaining the 90º difference between the
>> orientations of their polarization axes as you rotate both. There will
>> likely be a best orientation for the pair of polarizers.
>>
>> Alternatively, you might employ a long pass filter in front of the camera
>> so that only the longer emissions are passed while the shorter illumination
>> wavelengths are blocked. This presumes that you are illuminating with a
>> range of wavelengths shorter than those you wish to capture, otherwise some
>> of the illumination may be passed by the filter.
>>
>> Hope this helps,
>>
>> Bill Christens-Barry
>>
>> --
>> 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

[Robert Dougherty]

Anu,

As Herbie noted, there appears to be a shading issue.  You might try Polynomial Shading Corrector, http://www.optinav.info/Polynomial_Shading_Corrector.htm <http://www.optinav.info/Polynomial_Shading_Corrector.htm>

Bob

> On Sep 30, 2017, at 11:20 AM, anusuya pal <[hidden email]> wrote:
>
> Hello
>
> Here is the sample image (raw file) attached. I am using a polarizing
> microscope. And I think for my other samples, BaSiC is working good. But
> for particularly this type of image, I am not getting the background
> intensity and sample intensity differently. The range is more or less same.
> So, if I know any plugins that will solve this correction for the current
> image, It will be helpful.
>
> Thanks
> Anu
>
> On Sat, Sep 30, 2017 at 12:17 PM, Bill Christens-Barry <[hidden email]>
> wrote:
>
>> Anu,
>>
>> You mention wanting to reject reflected light; am I right in thinking that
>> the light you do want to capture in the image is due to fluorescence? If
>> so, two approaches come to mind:
>>
>> It might be possible to use linear polarizers. Reflected light often
>> retains the same linear polarization as that of the illumination, while
>> fluorescence emission is often highly depolarized. If you place a linear
>> polarizer in front of the light source and place another in front of the
>> camera, with its polarization axis perpendicular to that of the
>> illumination polarizer, the reflected light will be greatly attenuated
>> while the fluorescence emission will be decreased to a lesser extent (~ 2x
>> for full depolarization). This is especially true if the reflected light
>> has been reflected from the surface alone; translucent materials will allow
>> more penetration and consequent depolarization due to internal scattering,
>> which reduces the effectiveness of this approach. Try rotating the
>> illumination polarizer and the camera polarizer about their polarization
>> axes together, i.e. by maintaining the 90º difference between the
>> orientations of their polarization axes as you rotate both. There will
>> likely be a best orientation for the pair of polarizers.
>>
>> Alternatively, you might employ a long pass filter in front of the camera
>> so that only the longer emissions are passed while the shorter illumination
>> wavelengths are blocked. This presumes that you are illuminating with a
>> range of wavelengths shorter than those you wish to capture, otherwise some
>> of the illumination may be passed by the filter.
>>
>> Hope this helps,
>>
>> Bill Christens-Barry
>>
>> --
>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> <0 or 360.tif>

Robert P. Dougherty
President
OptiNav, Inc.
1414 127th Pl NE #106
Bellevue, WA 98005
[phone]
FAX [phone]
[hidden email]
www.optinav. com


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

[anusuya pal]

Hello,

Sorry for the late reply. As Herbie correctly pointed out, there is a
gradient in the background, and I want to remove that. The right side of
the image has more non-uniformity, which needs to be corrected. And as I
mentioned before, the background (shading) needs to be corrected properly
so that we don't lose any information about the sample but also, I can know
the mean intensity of the sample region properly. If I do without any
correction, it happens that the sample intensity is less than that of the
background, which is impossible as per the logic. The whole object is big,
the portion is taken with the background (which is the microscopic glass)
and the objective used is 5X.

Thanks for the suggestion Robert. I will check that shading corrector.


Thanks
Anu

On Sun, Oct 1, 2017 at 10:46 PM, Robert Dougherty <[hidden email]> wrote:

> Anu,
>
> As Herbie noted, there appears to be a shading issue.  You might try
> Polynomial Shading Corrector, http://www.optinav.info/
> Polynomial_Shading_Corrector.htm <http://www.optinav.info/
> Polynomial_Shading_Corrector.htm>
>
> Bob
>
> > On Sep 30, 2017, at 11:20 AM, anusuya pal <[hidden email]>
> wrote:
> >
> > Hello
> >
> > Here is the sample image (raw file) attached. I am using a polarizing
> > microscope. And I think for my other samples, BaSiC is working good. But
> > for particularly this type of image, I am not getting the background
> > intensity and sample intensity differently. The range is more or less
> same.
> > So, if I know any plugins that will solve this correction for the current
> > image, It will be helpful.
> >
> > Thanks
> > Anu
> >
> > On Sat, Sep 30, 2017 at 12:17 PM, Bill Christens-Barry <
> [hidden email]>
> > wrote:
> >
> >> Anu,
> >>
> >> You mention wanting to reject reflected light; am I right in thinking
> that
> >> the light you do want to capture in the image is due to fluorescence? If
> >> so, two approaches come to mind:
> >>
> >> It might be possible to use linear polarizers. Reflected light often
> >> retains the same linear polarization as that of the illumination, while
> >> fluorescence emission is often highly depolarized. If you place a linear
> >> polarizer in front of the light source and place another in front of the
> >> camera, with its polarization axis perpendicular to that of the
> >> illumination polarizer, the reflected light will be greatly attenuated
> >> while the fluorescence emission will be decreased to a lesser extent (~
> 2x
> >> for full depolarization). This is especially true if the reflected light
> >> has been reflected from the surface alone; translucent materials will
> allow
> >> more penetration and consequent depolarization due to internal
> scattering,
> >> which reduces the effectiveness of this approach. Try rotating the
> >> illumination polarizer and the camera polarizer about their polarization
> >> axes together, i.e. by maintaining the 90º difference between the
> >> orientations of their polarization axes as you rotate both. There will
> >> likely be a best orientation for the pair of polarizers.
> >>
> >> Alternatively, you might employ a long pass filter in front of the
> camera
> >> so that only the longer emissions are passed while the shorter
> illumination
> >> wavelengths are blocked. This presumes that you are illuminating with a
> >> range of wavelengths shorter than those you wish to capture, otherwise
> some
> >> of the illumination may be passed by the filter.
> >>
> >> Hope this helps,
> >>
> >> Bill Christens-Barry
> >>
> >> --
> >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >>
> >
> > --
> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> > <0 or 360.tif>
>
> Robert P. Dougherty
> President
> OptiNav, Inc.
> 1414 127th Pl NE #106
> Bellevue, WA 98005
> [phone]
> FAX [phone]
> [hidden email]
> www.optinav. com
>
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

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

[Herbie]

Good day Anu,

you have to decide whether you improve image acquisition or use post hoc
image correction, e.g. Shading Correction as proposed by Robert Dougherty.

I would vote for better image acquisition.

My impression is that you work with transmitted light and uneven
illumination but you should also consider highly diffuse incident
illumination.
(For 5x magnification an inverted photographic macro lens system may be
sufficient.)

Good luck

Herbie

:::::::::::::::::::::::::::::::::::::::::
Am 02.10.17 um 05:30 schrieb anusuya pal:

> Hello,
>
> Sorry for the late reply. As Herbie correctly pointed out, there is a
> gradient in the background, and I want to remove that. The right side of
> the image has more non-uniformity, which needs to be corrected. And as I
> mentioned before, the background (shading) needs to be corrected properly
> so that we don't lose any information about the sample but also, I can know
> the mean intensity of the sample region properly. If I do without any
> correction, it happens that the sample intensity is less than that of the
> background, which is impossible as per the logic. The whole object is big,
> the portion is taken with the background (which is the microscopic glass)
> and the objective used is 5X.
>
> Thanks for the suggestion Robert. I will check that shading corrector.
>
>
> Thanks
> Anu
>
> On Sun, Oct 1, 2017 at 10:46 PM, Robert Dougherty <[hidden email]> wrote:
>
>> Anu,
>>
>> As Herbie noted, there appears to be a shading issue.  You might try
>> Polynomial Shading Corrector, http://www.optinav.info/
>> Polynomial_Shading_Corrector.htm <http://www.optinav.info/
>> Polynomial_Shading_Corrector.htm>
>>
>> Bob
>>
>>> On Sep 30, 2017, at 11:20 AM, anusuya pal <[hidden email]>
>> wrote:
>>>
>>> Hello
>>>
>>> Here is the sample image (raw file) attached. I am using a polarizing
>>> microscope. And I think for my other samples, BaSiC is working good. But
>>> for particularly this type of image, I am not getting the background
>>> intensity and sample intensity differently. The range is more or less
>> same.
>>> So, if I know any plugins that will solve this correction for the current
>>> image, It will be helpful.
>>>
>>> Thanks
>>> Anu
>>>
>>> On Sat, Sep 30, 2017 at 12:17 PM, Bill Christens-Barry <
>> [hidden email]>
>>> wrote:
>>>
>>>> Anu,
>>>>
>>>> You mention wanting to reject reflected light; am I right in thinking
>> that
>>>> the light you do want to capture in the image is due to fluorescence? If
>>>> so, two approaches come to mind:
>>>>
>>>> It might be possible to use linear polarizers. Reflected light often
>>>> retains the same linear polarization as that of the illumination, while
>>>> fluorescence emission is often highly depolarized. If you place a linear
>>>> polarizer in front of the light source and place another in front of the
>>>> camera, with its polarization axis perpendicular to that of the
>>>> illumination polarizer, the reflected light will be greatly attenuated
>>>> while the fluorescence emission will be decreased to a lesser extent (~
>> 2x
>>>> for full depolarization). This is especially true if the reflected light
>>>> has been reflected from the surface alone; translucent materials will
>> allow
>>>> more penetration and consequent depolarization due to internal
>> scattering,
>>>> which reduces the effectiveness of this approach. Try rotating the
>>>> illumination polarizer and the camera polarizer about their polarization
>>>> axes together, i.e. by maintaining the 90º difference between the
>>>> orientations of their polarization axes as you rotate both. There will
>>>> likely be a best orientation for the pair of polarizers.
>>>>
>>>> Alternatively, you might employ a long pass filter in front of the
>> camera
>>>> so that only the longer emissions are passed while the shorter
>> illumination
>>>> wavelengths are blocked. This presumes that you are illuminating with a
>>>> range of wavelengths shorter than those you wish to capture, otherwise
>> some
>>>> of the illumination may be passed by the filter.
>>>>
>>>> Hope this helps,
>>>>
>>>> Bill Christens-Barry
>>>>
>>>> --
>>>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>>>
>>>
>>> --
>>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>> <0 or 360.tif>
>>
>> Robert P. Dougherty
>> President
>> OptiNav, Inc.
>> 1414 127th Pl NE #106
>> Bellevue, WA 98005
>> [phone]
>> FAX [phone]
>> [hidden email]
>> www.optinav. com
>>
>>
>> --
>> 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

[anusuya pal]

Thak you so much for the suggestions!
Thanks to Robert, this plugin removes all the artifacts!

Thanks
Anu



On Mon, Oct 2, 2017 at 5:14 AM, Herbie <[hidden email]> wrote:

> Good day Anu,
>
> you have to decide whether you improve image acquisition or use post hoc
> image correction, e.g. Shading Correction as proposed by Robert Dougherty.
>
> I would vote for better image acquisition.
>
> My impression is that you work with transmitted light and uneven
> illumination but you should also consider highly diffuse incident
> illumination.
> (For 5x magnification an inverted photographic macro lens system may be
> sufficient.)
>
> Good luck
>
> Herbie
>
> :::::::::::::::::::::::::::::::::::::::::
> Am 02.10.17 um 05:30 schrieb anusuya pal:
>
> Hello,
>>
>> Sorry for the late reply. As Herbie correctly pointed out, there is a
>> gradient in the background, and I want to remove that. The right side of
>> the image has more non-uniformity, which needs to be corrected. And as I
>> mentioned before, the background (shading) needs to be corrected properly
>> so that we don't lose any information about the sample but also, I can
>> know
>> the mean intensity of the sample region properly. If I do without any
>> correction, it happens that the sample intensity is less than that of the
>> background, which is impossible as per the logic. The whole object is big,
>> the portion is taken with the background (which is the microscopic glass)
>> and the objective used is 5X.
>>
>> Thanks for the suggestion Robert. I will check that shading corrector.
>>
>>
>> Thanks
>> Anu
>>
>> On Sun, Oct 1, 2017 at 10:46 PM, Robert Dougherty <[hidden email]>
>> wrote:
>>
>> Anu,
>>>
>>> As Herbie noted, there appears to be a shading issue.  You might try
>>> Polynomial Shading Corrector, http://www.optinav.info/
>>> Polynomial_Shading_Corrector.htm <http://www.optinav.info/
>>> Polynomial_Shading_Corrector.htm>
>>>
>>> Bob
>>>
>>> On Sep 30, 2017, at 11:20 AM, anusuya pal <[hidden email]>
>>>>
>>> wrote:
>>>
>>>>
>>>> Hello
>>>>
>>>> Here is the sample image (raw file) attached. I am using a polarizing
>>>> microscope. And I think for my other samples, BaSiC is working good. But
>>>> for particularly this type of image, I am not getting the background
>>>> intensity and sample intensity differently. The range is more or less
>>>>
>>> same.
>>>
>>>> So, if I know any plugins that will solve this correction for the
>>>> current
>>>> image, It will be helpful.
>>>>
>>>> Thanks
>>>> Anu
>>>>
>>>> On Sat, Sep 30, 2017 at 12:17 PM, Bill Christens-Barry <
>>>>
>>> [hidden email]>
>>>
>>>> wrote:
>>>>
>>>> Anu,
>>>>>
>>>>> You mention wanting to reject reflected light; am I right in thinking
>>>>>
>>>> that
>>>
>>>> the light you do want to capture in the image is due to fluorescence? If
>>>>> so, two approaches come to mind:
>>>>>
>>>>> It might be possible to use linear polarizers. Reflected light often
>>>>> retains the same linear polarization as that of the illumination, while
>>>>> fluorescence emission is often highly depolarized. If you place a
>>>>> linear
>>>>> polarizer in front of the light source and place another in front of
>>>>> the
>>>>> camera, with its polarization axis perpendicular to that of the
>>>>> illumination polarizer, the reflected light will be greatly attenuated
>>>>> while the fluorescence emission will be decreased to a lesser extent (~
>>>>>
>>>> 2x
>>>
>>>> for full depolarization). This is especially true if the reflected light
>>>>> has been reflected from the surface alone; translucent materials will
>>>>>
>>>> allow
>>>
>>>> more penetration and consequent depolarization due to internal
>>>>>
>>>> scattering,
>>>
>>>> which reduces the effectiveness of this approach. Try rotating the
>>>>> illumination polarizer and the camera polarizer about their
>>>>> polarization
>>>>> axes together, i.e. by maintaining the 90º difference between the
>>>>> orientations of their polarization axes as you rotate both. There will
>>>>> likely be a best orientation for the pair of polarizers.
>>>>>
>>>>> Alternatively, you might employ a long pass filter in front of the
>>>>>
>>>> camera
>>>
>>>> so that only the longer emissions are passed while the shorter
>>>>>
>>>> illumination
>>>
>>>> wavelengths are blocked. This presumes that you are illuminating with a
>>>>> range of wavelengths shorter than those you wish to capture, otherwise
>>>>>
>>>> some
>>>
>>>> of the illumination may be passed by the filter.
>>>>>
>>>>> Hope this helps,
>>>>>
>>>>> Bill Christens-Barry
>>>>>
>>>>> --
>>>>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>>>>
>>>>>
>>>> --
>>>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>>> <0 or 360.tif>
>>>>
>>>
>>> Robert P. Dougherty
>>> President
>>> OptiNav, Inc.
>>> 1414 127th Pl NE #106
>>> Bellevue, WA 98005
>>> [phone]
>>> FAX [phone]
>>> [hidden email]
>>> www.optinav. com
>>>
>>>
>>> --
>>> 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