Fusing Images: Color?

[Abhishek Sengupta]

When I am adding two images using ImageJ, one grayscale image (reference)
and another image pseudocolored green, the resultant image is entirely
green, i.e. the objects in the grayscale image are also colored to green. I
have tried pseudocoloring the reference image to blue but to no effect. Is
there a way to fuse Images iin ImageJ like a plugin that will overcome this
problem. The function that I was using was Process>Image Calculator>Add.

Thanks.

Abhishek Sengupta

[Albert Cardona]

The easy way:

- convert your grey 8-bit image to RGB (all 3 channels have the same
values, so no apparent change).
- convert your RGB image to RGB stack (separate channels)
- scroll to the green channel (the first slice)
- paste (with paste control "Add") your 8-bit green LUT image to the
green channel. The pasted green  LUT image is only visible as changes in
greyscale intensity, no color.
- convert the RGB stack to RGB: now the green color is visible.

Be sure to increase/decrease the brightness and contrast of the image to
paste, otherwise it may not be visible within the context of your target
image.

Albert

[Albert Cardona]

Correction: I meant "Max", not "Add", in the paste control menu.

The "Max" operation will choose the pixel with highest value from both
images.

You can test the paste interactively by selecting options in the Paste
control pull down menu.

Albert

[as538]

Thank you everyone for your responses. The images that I had were RGB images
and so I resolved it by using the RGB merge feature in ImageJ.

I really appreciated your quick responses and valuable suggestions.

Abhishek

On 6/10/07, [hidden email] <[hidden email]> wrote:

>
>
>
> *Forwarded Conversation*
> Subject: *Fusing Images: Color?*
> ------------------------
>
> * From: Abhishek Sengupta* <[hidden email]> To: ImageJInterest Group <
> [hidden email]>
> Date: Thu, Jun 7, 2007 at 8:08 PM
>
> When I am adding two images using ImageJ, one grayscale image (reference)
> and another image pseudocolored green, the resultant image is entirely
> green, i.e. the objects in the grayscale image are also colored to green.
> I have tried pseudocoloring the reference image to blue but to no effect. Is
> there a way to fuse Images iin ImageJ like a plugin that will overcome
> this problem. The function that I was using was Process>Image
> Calculator>Add.
>
> Thanks.
>
> Abhishek Sengupta
>
> --------
> * From: Rasband Wayne* <[hidden email]> To: Abhishek Sengupta <
> [hidden email]>
> Date: Thu, Jun 7, 2007 at 9:12 PM
>
> Try  converting both images to RGB and then using the "Average"
> operator in the Image Calculator.
>
> -wayne
> [Quoted text hidden]
>
> --------
> * From: F Javier Diez Guerra* <[hidden email]> Reply-To: ImageJInterest Group <
> [hidden email]>
> To: [hidden email]
> Date: Fri, Jun 8, 2007 at 2:48 AM
>
> Have you tried the "RGB Gray Merge" plugin?
>
> At 02:08 08/06/2007, you wrote:
> >When I am adding two images using ImageJ, one grayscale image (reference)
> >and another image pseudocolored green, the resultant image is entirely
> >green, i.e. the objects in the grayscale image are also colored to green.
> I
> >have tried pseudocoloring the reference image to blue but to no effect.
> Is
> >there a way to fuse Images iin ImageJ like a plugin that will overcome
> this
> >problem. The function that I was using was Process>Image Calculator>Add.
> >
> >Thanks.
> >
> >Abhishek Sengupta
>
> F Javier Diez-Guerra, PhD
> Profesor Titular
> Centro de Biologia Molecular Severo Ochoa
> Facultad de Ciencias, Universidad Autónoma
> Ctra Colmenar Viejo Km 15
> Cantoblanco, 28049 Madrid
> SPAIN
>
> phone:  [phone]
> Fax:      [phone]
> e-mail: [hidden email]
>
> --------
> * From: Nasim* <[hidden email]> Reply-To: ImageJ Interest Group <
> [hidden email]>
> To: [hidden email]
> Date: Fri, Jun 8, 2007 at 3:04 PM
>
> Abhishek,
>
>  You could also try downloading the MRIanalysisPak and using the 'Image
> overlay' there.
>
>  Nasim
> [Quoted text hidden]
> ---------------------------------
> Choose the right car based on your needs.  Check out Yahoo! Autos new Car
> Finder tool.
>
> --------
> * From: Albert Cardona* <[hidden email]> Reply-To: ImageJInterest Group <
> [hidden email]>
> To: [hidden email]
> Date: Sat, Jun 9, 2007 at 2:58 PM
>
> The easy way:
>
> - convert your grey 8-bit image to RGB (all 3 channels have the same
> values, so no apparent change).
> - convert your RGB image to RGB stack (separate channels)
> - scroll to the green channel (the first slice)
> - paste (with paste control "Add") your 8-bit green LUT image to the
> green channel. The pasted green  LUT image is only visible as changes in
> greyscale intensity, no color.
> - convert the RGB stack to RGB: now the green color is visible.
>
> Be sure to increase/decrease the brightness and contrast of the image to
> paste, otherwise it may not be visible within the context of your target
> image.
>
> Albert
>
> --------
> * From: Albert Cardona* <[hidden email]> Reply-To: ImageJInterest Group <
> [hidden email]>
> To: [hidden email]
> Date: Sat, Jun 9, 2007 at 5:07 PM
>
> Correction: I meant "Max", not "Add", in the paste control menu.
>
> The "Max" operation will choose the pixel with highest value from both
> images.
>
> You can test the paste interactively by selecting options in the Paste
> control pull down menu.
>
> Albert
>
> --------
>

--
Abhishek Sengupta
Drexel University
[hidden email]

[Nicholas Verge]

Abhishek,

There are literally dozens of algorithms for fusing greyscale  
single-channel images with multichannel colour iamges. It is a very  
important image processing procedure in remote sensing. There they are  
used to to produce a high resolution multispectral image, by fusing a wide  
spectral band, high resolution greyscale image with a colour-multispectral  
image of lower resolution. The wide variety of available methods reflects  
attempts to insert the high frequency information of the greuscale image  
into minimise into the nelarged colour image, and minimising the problem  
of colour distortion in high resolution colour product.

Note. The resolution enhancement aspect of these algorithms, is only a  
conventional resampling of the the colour image to the resolution of the  
greyscale image. So if your images are have both the same height and width  
as measured in pixels, ignore.

Try searching google using the terms "pan-sharpening" or "image fusion" or  
"resolution enhancment"

Common methods in use are:

Additive (alpha transparency merge)
HSI/HSL intensity/lightness substitution - replace the intsnsity/lightness  
componet of the enlarged low resolution colour image with the greyscale  
image (histogram matched)
PCA substitution - replace PC1 of the enlarged colour image with the  
greyscale image, then perform inverse PCA transform
Smoothed frequency intensity modulation
Various wavelet-transform-based fusion methods
High-pass filter method
Multi-resolution analysis intensity modulation.(MRAIM)
Pixel block intensity modulation
Brovey (chromaticity transform) method

Hope this may help.



----------------------------------------------------------------------
Nicholas J. Verge BSc. FGS