Image analysis

[Mike Tighe]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

?This is more of an ImageJ question. I am trying to quantify total area of inflamed vs normal lung tissue within a section. The problem is that the sections can have more or less airway-area and may even be inflated differently. I would like to remove the "empty space" or the background to make the numbers more accurate. I can measure area just fine but how could I remove the background from those measurements?


Thanks for any help!!

Mike

[JOEL B. SHEFFIELD]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Can you post an image or two?



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 Wed, Oct 29, 2014 at 3:55 PM, Mike Tighe <[hidden email]>
wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> ?This is more of an ImageJ question. I am trying to quantify total area of
> inflamed vs normal lung tissue within a section. The problem is that the
> sections can have more or less airway-area and may even be inflated
> differently. I would like to remove the "empty space" or the background to
> make the numbers more accurate. I can measure area just fine but how could
> I remove the background from those measurements?
>
>
> Thanks for any help!!
>
> Mike
>

[Martin Wessendorf-2]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Dear Dr. Tighe--

My guess is that you could subtract out the background by defining the
area that's "empty" using thresholding, saving that area as a
region-of-interest (ROI), and then inverting that selection so that you
then quantify everything that is NOT "empty".

However, it may be worth your time to ask the ImageJ listserv--it's
active and helpful.

Good luck!

Martin Wessendorf




On 10/29/2014 2:55 PM, Mike Tighe wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> ?This is more of an ImageJ question. I am trying to quantify total area of inflamed vs normal lung tissue within a section. The problem is that the sections can have more or less airway-area and may even be inflated differently. I would like to remove the "empty space" or the background to make the numbers more accurate. I can measure area just fine but how could I remove the background from those measurements?
>
>
> Thanks for any help!!
>
> Mike

--
Martin Wessendorf, Ph.D.                   office: [phone]
Assoc Prof, Dept Neuroscience                 lab: [phone]
University of Minnesota             Preferred FAX: [phone]
6-145 Jackson Hall, 321 Church St. SE    Dept Fax: [phone]
Minneapolis, MN  55455                    e-mail: [hidden email]

[Luis Farias]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Dear,
You can try
http://imagejconf.tudor.lu/program/poster/birgit_moeller1063026085
There is a link to MITOBO plugin for ImageJ
http://www2.informatik.uni-halle.de/agprbio/mitobo/index.php/Downloads

It really helped us.

cya


Luis Henrique Seabra de Farias*, Msc.*
E-mail: [hidden email] <[hidden email]>
Universidade Federal do Pará
Instituto de Ciências Biológicas
Laboratório de Parasitologia/Laboratório de Biologia Estrutural
Telefone: 913201-7927/913201-8232

2014-10-29 18:44 GMT-03:00 Martin Wessendorf <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Dear Dr. Tighe--
>
> My guess is that you could subtract out the background by defining the
> area that's "empty" using thresholding, saving that area as a
> region-of-interest (ROI), and then inverting that selection so that you
> then quantify everything that is NOT "empty".
>
> However, it may be worth your time to ask the ImageJ listserv--it's active
> and helpful.
>
> Good luck!
>
> Martin Wessendorf
>
>
>
>
>
> On 10/29/2014 2:55 PM, Mike Tighe wrote:
>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> Post images on http://www.imgur.com and include the link in your posting.
>> *****
>>
>> ?This is more of an ImageJ question. I am trying to quantify total area
>> of inflamed vs normal lung tissue within a section. The problem is that the
>> sections can have more or less airway-area and may even be inflated
>> differently. I would like to remove the "empty space" or the background to
>> make the numbers more accurate. I can measure area just fine but how could
>> I remove the background from those measurements?
>>
>>
>> Thanks for any help!!
>>
>> Mike
>>
>
> --
> Martin Wessendorf, Ph.D.                   office: [phone]
> Assoc Prof, Dept Neuroscience                 lab: [phone]
> University of Minnesota             Preferred FAX: [phone]
> 6-145 Jackson Hall, 321 Church St. SE    Dept Fax: [phone]
> Minneapolis, MN  55455                    e-mail: [hidden email]
>

[Keith Morris]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi Martin,

Being an inhalation toxicologist in a former research life, thresholding (ie. detecting the darker underlying tissue and ignoring the brighter air spaces) is normally the way to go. You threshold to get an overlying binary image of the detected tissue (typically red but you can select any colour), and the image analysis program can then be told to only measure tissue under this binary overlay image (i.e. the red detected/thresholded areas). Problems arise if your transmission light tissue image is poorly illuminated with dark shadows (I'm assuming it's a standard stained section in brightfield and not a fluorescence one). The dark shadows may give rise to false detection of air spaces when thresholding. Ideally to remove these take a shadow corrected image (an area with no tissue) and use this to background correct the image you capture (which will correct shadowing/shading across the image). Most image acquisition programs should have this option (we use NISElements BR).

Also make sure Kohler illumination is set up on the microscope before hand (ie. align the transmission halogen bulb correctly over the focused image for maximum brightness and minimum shadow). If the section is stained don't use phase contrast (phase contrast images of unstained tissue will be more difficult to threshold). Finally you may well have to edit the binary (detected/thresholded) image to remove false detection areas (dark shadows in the airspaces). Unfortunately modern day image analysis programmers have forgotten the merits of a binary image (thresholded) image editor, but that's no problem just edit a copy of the image in Photohshop/Elements or GIMP and turn these dark air space problem areas into pure white (draw region and fill with white). Most image analysis programs allow copy and paste to and from Photoshop to speed this external image editing up.  These white areas will then be ignored when you set thresholding for the dark tissue. Hopefully with background corrected capture and Kohler setup correctly you won't have to do much editing, but there will still be crud, hairs etc. you don't want measured.  Where the image has an outer edge (at the lung periphery) you can use a region tool to ensure only the tissue area + spaces is selected for thresholding/measurement.

As Luis suggests, you may find utilities that do the thresholding/detecting and measurement for you, but you still have to get the microscope image right (i.e. evenly illuminated or edited) for the tissue to be detected & measured (analysed) correctly. You may want to measure the entire tissue (including inner air spaces) and then measure the tissue only (no air spaces). The ratio of the two (um2) values will give the relative increase in tissue 'thickness' and reduction of air spaces with the fibrotic lungs in a given area of lung. Inflated lung differences will be real problem though so you must ensure they all inflated at the same pressure height of fixative, otherwise collapsed healthy lung could look and measure worse than diseased lung (you may be able to work out some sort correction in this case though). As Martin suggests also ask the ImageJ listserver for ImageJ specific tool tips and the microscopy.com listserver may offer further lung fixation/image analysis ideas (although many of here are on both servers).

Hope this helps.

Regards

Keith

-----------------------------------------------------------
Dr Keith J Morris
Cellular Imaging Microscopy Core,
The Wellcome Trust Centre for Human Genetics,
Roosevelt Drive,
Oxford,
OX3 7BN,
United Kingdom.
 
Tel:  +44   ( 0 ) 1865  287568
Email:   [hidden email]
Webpage: www.well.ox.ac.uk/microscopy-facilities
 


-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Luis Farias
Sent: 29 October 2014 23:38
To: [hidden email]
Subject: Re: Image analysis

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Dear,
You can try
http://imagejconf.tudor.lu/program/poster/birgit_moeller1063026085
There is a link to MITOBO plugin for ImageJ http://www2.informatik.uni-halle.de/agprbio/mitobo/index.php/Downloads

It really helped us.

cya


Luis Henrique Seabra de Farias*, Msc.*
E-mail: [hidden email] <[hidden email]> Universidade Federal do Pará Instituto de Ciências Biológicas Laboratório de Parasitologia/Laboratório de Biologia Estrutural
Telefone: 913201-7927/913201-8232

2014-10-29 18:44 GMT-03:00 Martin Wessendorf <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Dear Dr. Tighe--
>
> My guess is that you could subtract out the background by defining the
> area that's "empty" using thresholding, saving that area as a
> region-of-interest (ROI), and then inverting that selection so that
> you then quantify everything that is NOT "empty".
>
> However, it may be worth your time to ask the ImageJ listserv--it's
> active and helpful.
>
> Good luck!
>
> Martin Wessendorf
>
>
>
>
>
> On 10/29/2014 2:55 PM, Mike Tighe wrote:
>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> Post images on http://www.imgur.com and include the link in your posting.
>> *****
>>
>> ?This is more of an ImageJ question. I am trying to quantify total
>> area of inflamed vs normal lung tissue within a section. The problem
>> is that the sections can have more or less airway-area and may even
>> be inflated differently. I would like to remove the "empty space" or
>> the background to make the numbers more accurate. I can measure area
>> just fine but how could I remove the background from those measurements?
>>
>>
>> Thanks for any help!!
>>
>> Mike
>>
>
> --
> Martin Wessendorf, Ph.D.                   office: [phone]
> Assoc Prof, Dept Neuroscience                 lab: [phone]
> University of Minnesota             Preferred FAX: [phone]
> 6-145 Jackson Hall, 321 Church St. SE    Dept Fax: [phone]
> Minneapolis, MN  55455                    e-mail: [hidden email]
>

[Sathya]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****



Hi Mike,

I would recommend to use
stereological estimation to perform area/ area fraction analysis. When it comes
to lung tissue, the air volume within the tissue is a major problem and has to
be taken into consideration. The air space varies for each field of view and
the entire region of interest should be considered while measuring area. A
systematic uniform sampling should be obtained throughout the tissue.  Just to give a few articles on lung tissue
stereology and how it can be performed:

-        
Jan Philipp Schneider, Matthias Ochs. Chapter
12 - Stereology of the Lung. Pages 257-294. In: Laboratory Methods in Cell
Biology Imaging Volume 113, Pages 1-414 (2013);  Edited by P. Michael Conn

-        
Ochs M, Mühlfeld C. Quantitative
microscopy of the lung - a problem-based approach Part 1: Basic principles of
lung stereology. Am J Physiol Lung Cell Mol Physiol. 2013 Apr 26.

 To deal with the
imaging problems as suggested by Keith (like linearity, brightness and
uniformity in imaging from one image with another), I had the opportunity to
attend a few webinars and also discuss about using the ChromaCal color
calibration slide (from Datacolor) to perform consistent and uniform calibration
for bright field images. For more information about what ChromaCal can do,
please visit http://scientific.datacolor.com/
(No commercial interest).

 If you require more
information about the methods and performing lung tissue stereology, please send
a mail and I can provide you with details. Good luck.

 Sathya Srinivasan

Manager

RUN Advanced Optical
Microscopy Core Facility

University of Calgary

Calgary, Alberta, Canada
 

> Date: Wed, 29 Oct 2014 19:55:08 +0000
> From: [hidden email]
> Subject: Image analysis
> To: [hidden email]
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> ?This is more of an ImageJ question. I am trying to quantify total area of inflamed vs normal lung tissue within a section. The problem is that the sections can have more or less airway-area and may even be inflated differently. I would like to remove the "empty space" or the background to make the numbers more accurate. I can measure area just fine but how could I remove the background from those measurements?
>
>
> Thanks for any help!!
>
> Mike