Extract Color from Grayscale Image not Index

Hello. I am trying to extract the color value from a grayscale image but my program currently gives the indexes. For example for color white, I get 0.53, for black it's 0.16. The imtool displays the actual colors I want to extract(example pic is attached). My code is pretty simple as shown below.

colors = image(:);
uniqueColors = unique(colors);

Any help would be appreciated, thanks!

[Adam]

The colour is presumably just a normalisation over the full data range. I don't use imtool so I don't know how it scales the data. There is no such thing as the 'actual colour' as such though - it is all dependent on how the data is scaled. If you are happy with the way it is scaled in the picture than you just need to normalise your data from its initial range to a 0-1 range (and make a triplet out of the results for greyscale colour).

I do that using a custom class I have but there are functions in the image processing toolbox I think that Image Analyst will now better than me - I forget to use most of the convenience functions in there.

[Jason Bond]

Thank you Adam. In order to normalise between 0 and 1, I added the following line:

colorsN = (colors - min(colors)) / ( max(colors) - min(colors));

And this does give values between 0 and 1. But when I compare these values, especially for gray colors, the indexes are off by some degree of accuracy. For example, colorsN(2) is 0.6486, but the value shown on the image tool is 0.651. How can I find the exact match? Thank you!

Answer 1

[Image Analyst]

Just index your image

grayLevel = yourImage(row, column);

[Jason Bond]

Thank you for your reply Image Analyst. I have tried looping through each pixel, i.e. image(i,j) but this still returns the index and not the color that is displayed on the imtool figure. For example, when the color is black, image(i,j) returns 0.16 but I need this to be 0. When the color is white, image(i,j) returns 0.53 but I need this to be 1. Thanks!

[Image Analyst]

It's returning the gray level, not the class number. So if you want class numbers, so that 0.16 gray level = class #0, 0.53 = class #1, and 0.87 = class #2, and so on, then you're going to have to find the unique gray levels in the image and convert them to a class number.

uniqueGrayLevels = unique(yourImage(:));

Then use imquantize(), something like

quant_A = imquantize(yourImage, uniqueGrayLevels )

Now the values of quant_A will be 1,2,3 etc. instead of the fractional numbers.

[Jason Bond]

Thanks for all your help.