How can i convert an Indexed Image To a Binary Image ?

Hello

How can i convert the blue into white and the other colours into black ?

The image is attached.

Answer 1

[DGM]

The attached image is a JPG; as such, it is not an indexed image. So we now have two interpretations of the question:

"The image is actually an indexed image, and I want one specific color to be the foreground"

[inpict,CT] = imread('op1.png'); % an actual indexed image
imshow(inpict,CT)

% select the medium purple regions
fgcolor = 2; % the index of the chosen color
mask = inpict == fgcolor; % an exact selection
imshow(mask)

●

"The image is a lossy pseudocolor image, and I want one specific color to be the foreground"

inpict = imread('op1.jpg'); % a low-quality JPG of the same image
imshow(inpict)

% an exact selection won't work with all the error
fgcolor = [152 58 143]; % the chosen color
mask = all(inpict == permute(fgcolor,[1 3 2]),3); % an exact selection
imshow(mask)

% try using a tolerance
tol = 0.1; % pick some tolerance
mask = all(abs(im2double(inpict) - permute(fgcolor/255,[1 3 2])) < tol,3);
imshow(mask)

[Walter Roberson]

As the original poster asks about converting "blue" into white, my interpretation is that they want all of the shades of blue to be turned into white. It is not immediately clear whether they want the intensity of the white to be equal to the intensity of the corresponding blue, or if they want all shades of blue to come out as the same shade of white.

In order to convert all shades of blue, I would suggest converting the RGB version of the image into HSV, and then look for locations where the hue is in the range 0.667 +/- 0.0625 (or so). The exact hue range to select is a bit uncertain -- how close to cyan do you want to allow, and how close to purple do you want to allow?

If you did want the shade of white to be the same as the shade of blue, then I would construct an HSV array with Hue and Saturation all 0, and with Value the same as corresponding "blue" pixels. Indeed if you were to do that and also set the Value for the unselected pixels to 0, then hsv2rgb would give the final result.

TheRGBImage = imread('ductile.JPG');

HSV = rgb2hsv(TheRGBImage);
H = HSV(:,:,1);
mask = (2/3 - 1/16) <= H & H <= (2/3 + 1/16);
S = HSV(:,:,3);
NewS = zeros(size(S),'like',S);
NewS(mask) = S(mask);
NewHSV = cat(3, zeros(size(HSV,1),size(HSV,2),2), NewS);
newRGB = hsv2rgb(NewHSV);

subplot(2,1,1); imshow(TheRGBImage);
subplot(2,1,2); imshow(newRGB);

●