insertShape function doesn't read color triplet

The function is able to read colors as strings ('red', 'green'...etc...) but cannot read modified color vectors like [1 0 0] or [0.6 1 0.2]. If I use this as an input the shape just come out black.

My color triplet is type double.

Here is the line:

img = insertShape(img, 'Line', position(i,:), 'Color', ColMap(c,:), 'LineWidth', w);

position and line width are perfect. ColMap ia a Mx3 matrix double.

What I dont understand is that using the function 'plot' to simply plot a line over an image everything works perfectly including the color. Meaning, the following line works:

line = plot([x1 x2], [y1 y2], 'Color', ColMap(c,:), 'LineWidth', w);

Any idea?

thank you

Answer 1

[Image Analyst]

Unlike other functions that require colors in the range 0-1, insertShape expects values in the range 0-255: Try this:

% Try it with a gray scale image.
img = imread('moon.tif');
position = [50, 20, 200, 500];
i = 1;
w = 13;
c = 220;
ColMap = jet(256); % In the range 0-1
customColor = round(255 * ColMap(c, :))
img = insertShape(img, 'Line', position(i,:), 'Color', customColor, 'LineWidth', w);
% img is now converted into an RGB image.  No longer a gray scale image.
subplot(2, 1, 1);
imshow(img);
axis('on', 'image');

% Try it with a color image.
img = imread('peppers.png');
img = insertShape(img, 'Line', position(i,:), 'Color', customColor, 'LineWidth', w);
subplot(2, 1, 2);
imshow(img);
axis('on', 'image');

[gabriele]

Correct. wow that was easy, should have thought about that.

Thank you!

[DGM]

It appears that the documentation is simply wrong -- rather, the webdocs are wrong, but the synopsis is correct.

It will accept unit-scale color specification, but only so long as the image is floating point.

position = [50, 20, 200, 500];
color = lines(1); % a very familiar blue (unit-scale double)
img = im2double(imread('peppers.png')); % double
img = insertShape(img, 'Line', position, 'Color', color, 'LineWidth',15);
imshow(img);

color = double(im2uint16(color)); % uint16-scale float
img = im2uint16(imread('peppers.png')); % uint16
img = insertShape(img, 'Line', position, 'Color',color, 'LineWidth',15);
imshow(img);

color = int16(im2uint8(uint16(color))); % uint8-scale int16
img = im2uint8(imread('peppers.png')); % uint8
img = insertShape(img, 'Line', position, 'Color',color, 'LineWidth',15);
imshow(img);

In other words, the color tuple must be scaled to match the class of the image, regardless of the class of the tuple itself.

See also:

https://www.mathworks.com/matlabcentral/answers/390090-can-someone-help-me-with-changing-colors-on-insertshape-function#answer_1307451