I'm using imagesc to display data in grayscale. I want to be able to set the color for one particular value to red. How do I accomplish this?

I have a 2-d matrix of type double. I want to display the data while retaining the values in the image. The image is to be displayed in gray scale, except for a particular value which I'd like to show up in red. This value is not fixed and can be changed later. I want to use imagesc for this (and change the limits, i.e. imagesc(data,[min max]),colormap('gray') and I only want those pixels with the exact value to be red. Is there a way to do this?

[Jason]

I should have mentioned that I want to do this programmatically, not using the colormapeditor.

Answer 1

[Image Analyst]

Try using a colormap

cmap = gray(256);
% Set one gray level, 8, to red.
% Remember, row #1 is 0 gray levels.
% so row #9 is 8 gray levels.
cmap(9,:) = [1,0,0];
colormap(cmap); % Apply the custom colormap.
colorbar;

[Jason]

Hi Image Analyst!

The problem I have with this is as it's written is how do I figure out which cmap row to edit and wouldn't this potentially create a red color for pixels other than what I want? What I mean is, if I only want to highlight in red a pixel with a specific value and not a very small spread of pixel values. I hope that makes sense.

[Image Analyst]

I showed you. Did you not see how I made it so that only gray level 8 would be in red by setting row #9 of the color map to red (1,0,0)?

What gray level do you want to assign? Let's say it's n, then set row (n+1) to [1,0,0].

Tell me what your gray level is and I'll tell you the line of code.

[Jason]

Oh I'm sorry, I should have included another piece of information. The matrix I want to display can be 128x128 and contain more than 2^14 unique data values. The values are real numbers with many digits to the right of the decimal. Since the gray map you have in your example only contains 256 levels, the likely hood that gray level 8 is the exact value I'm looking for is rare (unless I'm misunderstanding something).

With that in mind, here's a code snippet

a = -10;
b = 10;
data = (b-a).*rand(128,128) + a;
desiredValue = data(33,33);  % picking a value to modify to red

I only want values that are equal to desiredValue to show in red.

I thought if I made a custom colormap of the size of the unique values in data, i.e. if there are 16284 unique values, do

uData = unique(data);
cmap = gray(numel(uData);
idx = find(uData == desiredValue);
cmap(idx) = [1, 0, 0];
imagesc(data),colormap(cmap),colorbar
would do what I want, but it didn't highlight in red the pixel at loc 33,33

I hope this makes it better to understand what I'm trying to do.

[Image Analyst]

If you want 8 in a floating point image, it will be very very rare indeed. See the FAQ: http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F

You're going to have to set a tolerance and find pixels in that tolerance band. This might be easier to do if you create an RGB image than use a pseudocolor lookup table (colormap) on an indexed image.

GL1 = 7.995;
GL2 = 8.005;
binaryImage = grayImage >= GL1 & grayImage <= GL2;
% Extract the individual red, green, and blue color channels.
redChannel = grayImage ;
greenChannel = grayImage ;
blueChannel = grayImage ;
maxGL = max(grayImage(:));
redChannel(binaryImage) = maxGL;
greenChannel(binaryImage) = 0;
blueChannel(binaryImage) = 0;
% Recombine separate color channels into a single, true color RGB image.
rgbImage = cat(3, redChannel, greenChannel, blueChannel);
rgbImage = rgbImage / maxGL; % Scale to -1 range for display.
imshow(rgbImage);

That's untested code just off the top of my head.