ImageJ

Measuring of sponge spicule microscopy images

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

8 messages Options

eburell

Measuring of sponge spicule microscopy images

Dear members,

I am currently attempting to use ImageJ to measure the length and width of spicules found in sponges as a part of an assigned project. Unfortunately I am new to the usage of ImageJ and image analysis in general. I have two questions regarding ImageJ and image analysis.

The main problem I am facing right now is that I am unable to, more or less automatically, create a useful ROI from an microscopy image of a spicule sample (automation being part of the assignment). The ROI I'm trying to create is an an outline or a skeleton of the spicule for measurement. Attempts to skeletonize or otherwise create a binary image from the source image has been unsuccessful partly because of a poor source image and partly because of little to no knowledge regarding these operations (though I suspect the nature of the spicule images might also contribute). I have noticed that using a skeleton-based method makes the measurement of spicule thickness impossible and I am unsure of how I should counter this issue.

Am I approaching this measuring problem correctly?
Is there any recommended practice for this type of ROI-building and measuring?
Are there any tutorials or other documentation that detail a similar measuring processes for spike-shapes?

A sample microscopy image from the set i was given can be seen here: (https://skitch.com/aiden/r7j81/img-7369)
A rough sketch of the ROI I'm attempting to build can be seen here marked in red: (https://skitch.com/aiden/r7j86/img-7369-edited)

Regards,
Eli

Ayse Turak

Colorizing of binary images

Dear Members,

I am new to imagej and I was wondering if there was a way to colorize binary images -- either converting back into a psuedo-RGB image or to assign a different color value to the black portions, i.e. making all red

any advice would be great

Ayse

Arne Seitz

Re: Colorizing of binary images

Hi Ayse,

the easiest way to change the appearance of your (binary) image is to use Lookup tables (LUT).
You can find them under: Image->Lookup Tables.

Best regards
Arne

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Ayse Turak
Sent: mardi 3 mai 2011 16:18
To: [hidden email]
Subject: Colorizing of binary images

Dear Members,

I am new to imagej and I was wondering if there was a way to colorize binary images -- either converting back into a psuedo-RGB image or to assign a different color value to the black portions, i.e. making all red

any advice would be great

Ayse

Ayse Turak

Re: Colorizing of binary images

Dear Arne,

thanks for that -- my one problem is that no matter what I choose, black areas become coloured, but the areas that were once white now turn black (or coloured if I invert the image). Is there a way to have all the black pixels red for example, but keep the white pixels white?

sorry if this seems basic, I tried every variation of the LUT but none of them seem to work for me

regards

Ayse

On May 4, 2011, at 19:45 , Seitz Arne wrote:

> Hi Ayse,
>
> the easiest way to change the appearance of your (binary) image is to use Lookup tables (LUT).
> You can find them under: Image->Lookup Tables.
>
> Best regards
> Arne
>
> -----Original Message-----
> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Ayse Turak
> Sent: mardi 3 mai 2011 16:18
> To: [hidden email]
> Subject: Colorizing of binary images
>
> Dear Members,
>
> I am new to imagej and I was wondering if there was a way to colorize binary images -- either converting back into a psuedo-RGB image or to assign a different color value to the black portions, i.e. making all red
>
> any advice would be great
>
> Ayse
>

Dr. Ayse Turak <[hidden email]>
Visiting Assistant Professor
Faculty of Engineering and Natural Sciences
Sabanci University
Orhanli, Tuzla
34956 Istanbul, TURKEY
+90 216 483 9000 2126

Lim Soon Yew

Re: Colorizing of binary images

Hi Ayse,

You can use "Image>Color>Edit LUT" to create your own custom lookup table.

Best Regards,
John
________________________________________
From: ImageJ Interest Group [[hidden email]] On Behalf Of Ayse Turak [[hidden email]]
Sent: Wednesday, May 11, 2011 5:52 PM
To: [hidden email]
Subject: Re: Colorizing of binary images

Dear Arne,

thanks for that -- my one problem is that no matter what I choose, black areas become coloured, but the areas that were once white now turn black (or coloured if I invert the image). Is there a way to have all the black pixels red for example, but keep the white pixels white?

sorry if this seems basic, I tried every variation of the LUT but none of them seem to work for me

regards

Ayse

On May 4, 2011, at 19:45 , Seitz Arne wrote:

Dr. Ayse Turak <[hidden email]>
Visiting Assistant Professor
Faculty of Engineering and Natural Sciences
Sabanci University
Orhanli, Tuzla
34956 Istanbul, TURKEY
+90 216 483 9000 2126

David Strachan

Re: Colorizing of binary images

Hi Ayse,

If I understand you correctly you can start with your binary image(8-bit
type), make a new image the same size as this with all pixels set to
white(8-bit type), then use Image - Colour - Merge Channels function.
For the red channel select the all white image, for the green channel
select your original binary image, for the blue channel again select
your original binary image, grey channel select none.

And that should result in an RGB image that has white where there was
white and Red where there was black in your original binary image.

If you do not get this, check in case that the original binary image you
have is displayed with an inverted L.U.T.

Here is a small macro below to perform the task you want.

// Questions to D.Strachan(at)Beatson.gla.ac.uk
// Start with your binary image open
if (nImages!=1){
exit("You should have only your binary image open to start
with");
}
run("8-bit");// make sure that the binary image is an 8-bit type
imageName = getTitle();
print (imageName);
// Get the dimentions of this binary image
height = getHeight();
width = getWidth();
// Create a new image of the same size and bit depth
// with all pixels set to white
newImage("white", "8-bit White", width, height, 1);
//Create a new RGB image using the All White as Red channel and your
original image as the Green and Blue
//This means that what was white, stays white, and what was black goes
red.
run("Merge Channels...", "red=white green="+imageName+"
blue="+imageName+" gray=*None*");

regards

David Strachan
The Beatson Institute for Cancer Research
Glasgow G61 1BD
* E-mail [hidden email]

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of
Lim Soon Yew
Sent: 11 May 2011 11:11
To: [hidden email]
Subject: Re: Colorizing of binary images

Hi Ayse,

You can use "Image>Color>Edit LUT" to create your own custom lookup
table.

Best Regards,
John
________________________________________
From: ImageJ Interest Group [[hidden email]] On Behalf Of Ayse
Turak [[hidden email]]
Sent: Wednesday, May 11, 2011 5:52 PM
To: [hidden email]
Subject: Re: Colorizing of binary images

Dear Arne,

thanks for that -- my one problem is that no matter what I choose, black
areas become coloured, but the areas that were once white now turn black
(or coloured if I invert the image). Is there a way to have all the
black pixels red for example, but keep the white pixels white?

sorry if this seems basic, I tried every variation of the LUT but none
of them seem to work for me

regards

Ayse

On May 4, 2011, at 19:45 , Seitz Arne wrote:

> Hi Ayse,
>
> the easiest way to change the appearance of your (binary) image is to
use Lookup tables (LUT).

> You can find them under: Image->Lookup Tables.
>
> Best regards
> Arne
>
> -----Original Message-----
> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of
> Ayse Turak
> Sent: mardi 3 mai 2011 16:18
> To: [hidden email]
> Subject: Colorizing of binary images
>
> Dear Members,
>
> I am new to imagej and I was wondering if there was a way to colorize
> binary images -- either converting back into a psuedo-RGB image or to
> assign a different color value to the black portions, i.e. making all
> red
>
> any advice would be great
>
> Ayse
>

Dr. Ayse Turak <[hidden email]>
Visiting Assistant Professor
Faculty of Engineering and Natural Sciences Sabanci University Orhanli,
Tuzla
34956 Istanbul, TURKEY
+90 216 483 9000 2126

ali ayata

Re: Colorizing of binary images

In reply to this post by Ayse Turak

hi.
you can do this with photoshop #color replacement tool.

try this; http://www.digital-photography-school.com/changing-color-in-photoshop

ij de de mutlaka bunu yapmanın bir yolu vardır ama en kolayı bu bence.

--- On Wed, 5/11/11, Ayse Turak <[hidden email]> wrote:

From: Ayse Turak <[hidden email]>
Subject: Re: Colorizing of binary images
To: [hidden email]
Date: Wednesday, May 11, 2011, 5:52 AM

Dear Arne,

thanks for that -- my one problem is that no matter what I choose, black areas become coloured, but the areas that were once white now turn black (or coloured if I invert the image). Is there a way to have all the black pixels red for example, but keep the white pixels white?

sorry if this seems basic, I tried every variation of the LUT but none of them seem to work for me

regards

Ayse

On May 4, 2011, at 19:45 , Seitz Arne wrote:

Dr. Ayse Turak <[hidden email]>
Visiting Assistant Professor
Faculty of Engineering and Natural Sciences
Sabanci University
Orhanli, Tuzla
34956 Istanbul, TURKEY
+90 216 483 9000 2126

Krs5

Re: Colorizing of binary images

Hi Ayse,

Try:
Change type to RGB and use Image-->Color-->Split Channel
Fill the whole 'red' image with red (Select All-->Fill (with red selected as fill color))
Merge the three images again Image-->Color-->Merge Channels

And you should get a red image with white selections.

Hope it works

Kees

Dr Ir K.R. Straatman
Senior Experimental Officer
Centre for Core Biotechnology Services
College of Medicine, Biological Sciences and Psychology
http://www.le.ac.uk/biochem/microscopy/home.html
ImageJ workshops 21 and 22 July: http://www.le.ac.uk/biochem/microscopy/ImageJ2011.html

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of ali ayata
Sent: 11 May 2011 14:31
To: [hidden email]
Subject: Re: Colorizing of binary images

hi.
you can do this with photoshop #color replacement tool.

try this; http://www.digital-photography-school.com/changing-color-in-photoshop

ij de de mutlaka bunu yapmanın bir yolu vardır ama en kolayı bu bence.

--- On Wed, 5/11/11, Ayse Turak <[hidden email]> wrote:

From: Ayse Turak <[hidden email]>
Subject: Re: Colorizing of binary images
To: [hidden email]
Date: Wednesday, May 11, 2011, 5:52 AM

Dear Arne,

thanks for that -- my one problem is that no matter what I choose, black areas become coloured, but the areas that were once white now turn black (or coloured if I invert the image). Is there a way to have all the black pixels red for example, but keep the white pixels white?

sorry if this seems basic, I tried every variation of the LUT but none of them seem to work for me

regards

Ayse

On May 4, 2011, at 19:45 , Seitz Arne wrote:

Dr. Ayse Turak <[hidden email]>
Visiting Assistant Professor
Faculty of Engineering and Natural Sciences
Sabanci University
Orhanli, Tuzla
34956 Istanbul, TURKEY
+90 216 483 9000 2126