how to get the location and the color of a pixel in workspace by clicking with the mouse - Matlab

Hello,

I have some fiducial points in an image and I need the user to select the first point and get the coordinates of that point and the colour vector in workspace.

At the moment I have only found:

Data cursor, but it only gives the location, not the colour Impixel, apparently should give both, but it is a bit confusing and it's not working very well imroi only gives location too from what I've read

Can you please help me with this? Also, can the colour vector be in the Lab colorspace? (transform the image first, and then click in the point?)

Many thanks! Hector

Answer 1

[Chad Greene]

I wrote a colorpicker function to make this a bit easier.

Answer 2

[Image Analyst]

Use impixelinfo(), available in the Image Processing Toolbox. Instructions are in the help.

[HecLondon]

I found that it is possible to do what I wanted with the impixel function, so that's solved. Now I wonder: is it possible to change my photo to Lab colorspace and get the values of a and b with impixel instead of the rgb values? Also, I want to use the coordinates of the pixel to place the zero angle, to order my fiducial points in a clockwise direction. However atan2 always have the zero angle in the horizontal direction, and in my case it would be at a certain angle, is there any way to solve this? Many thanks!

[Image Analyst]

You can just get the value from the lab array directly:

labColor = lab(row, column, :);

I don't know what you're talking about with respect to angle - why there's an offset. Red is at zero degrees, yellow at 90, green at 180 and blue at 270 (like my avatar). Why would they not be? What is the purpose of an angular offset? If you want you can go to hsv colorspace with rgb2hsv() and h is the normalized hue angle.

[Image Analyst]

See this demo and you'll see the RGB and LAB are not the same.

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
imtool close all;  % Close all imtool figures if you have the Image Processing Toolbox.
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 20;

% Read in a standard MATLAB color demo image.
folder = fullfile(matlabroot, '\toolbox\images\imdemos');
baseFileName = 'peppers.png';
% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
if ~exist(fullFileName, 'file')
  % Didn't find it there.  Check the search path for it.
  fullFileName = baseFileName; % No path this time.
  if ~exist(fullFileName, 'file')
    % Still didn't find it.  Alert user.
    errorMessage = sprintf('Error: %s does not exist.', fullFileName);
    uiwait(warndlg(errorMessage));
    return;
  end
end
rgbImage = imread(fullFileName);
% Get the dimensions of the image.  numberOfColorBands should be = 3.
[rows, columns, numberOfColorBands] = size(rgbImage);
% Display the original color image.
imshow(rgbImage);
title('Original Color Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'units','normalized','outerposition',[0 0 1 1]);

promptMessage = sprintf('Please click a point in the image,\nor Quit to exit?');
titleBarCaption = 'Continue?';
button = questdlg(promptMessage, titleBarCaption, 'Continue', 'Quit', 'Continue');
if strcmpi(button, 'Quit')
  return;
end
[x,y] = ginput(1);
% Put a cross where they clicked.
hold on;
plot(x, y, 'w+', 'MarkerSize', 50);
% Get the location they click on.
row = int32(y);
column = int32(x);
% Extract the RGB color from that location.
rgbColor = rgbImage(row, column, :)
cform = makecform('srgb2lab');
% Get lab in range 0-255 for L.
labImage1 = applycform(rgbImage,cform);
% Get lab in range 0-100 for L
labImage2 = applycform(im2double(rgbImage),cform);
% Extract the RGB color from that location.
labColor1 = labImage1(row, column, :)
labColor2 = labImage2(row, column, :)
message = sprintf('Done!\n(x,y) = (%.3f, %.3f)\n(row, column) = (%d, %d)\n(R,G,B) = (%d, %d, %d)\n(L,A,B) = (%d, %d, %d)\n(L,A,B) = (%.3f, %.3f, %.3f)\n',...
  x,y,row,column, ...
  rgbImage(row, column, 1),  rgbImage(row, column, 2),  rgbImage(row, column, 3), ...
  labImage1(row, column, 1), labImage1(row, column, 2), labImage1(row, column, 3), ...
  labImage2(row, column, 1), labImage2(row, column, 2), labImage2(row, column, 3))
uiwait(helpdlg(message));

[HecLondon]

Many thanks and sorry for the delay.

When it says towards the end "Extract the RGB color from that location" did you mean the LAB?

I tried the demo. Previously, my color segmentation code was based in choosing an ROI within the fiducial point and getting the mean a and b values. I put the values obtained from your demo (range 0-255 as I didn't use im2double in my code) but the color segmentation is not working. Most of the code is from a demo I found, so I copy it here so that you can check:

    %load the photo and resize it to reduce the time of the code
fabric2 = imread('skin2.JPG');
fabric = imresize(fabric2, 0.2);

%Define the number of colors in the photo (we just need the skin and
%fiducial points, but it is possible to add more) and a matrix for the ROI
nColors = 2;

%This is to convert the RGB image into an L*a*b* image
cform = makecform('srgb2lab');
lab_fabric = applycform(fabric,cform);

%This is to calculate the mean 'a*' and 'b*' value for each area extracted with roipoly. These values serve as color markers in 'a*b*' space
a = lab_fabric(:,:,2);
b = lab_fabric(:,:,3);
color_markers = repmat(0, [nColors, 2]);

for count = 1:nColors% this is the part I changed, just putting the     %values obtained from your demo
color_markers(count,1) = 14 ;
color_markers(count,2) = -22 ;
end

%Create an array that contains your color labels
color_labels = 0:nColors-1;

%Initialize matrices to be used in the nearest neighbor classification
a = double(a);
b = double(b);
distance = repmat(0,[size(a), nColors]);

%Perform classification
for count = 1:nColors
distance(:,:,count) = ( (a - color_markers(count,1)).^2 + ...
                    (b - color_markers(count,2)).^2 ).^0.5;
end

[value, label] = min(distance,[],3);
label = color_labels(label);
clear value distance;

%Display Results of Nearest Neighbor Classification
rgb_label = repmat(label,[1 1 3]);
segmented_images = repmat(uint8(0),[size(fabric), nColors]);

for count = 1:nColors
color = fabric;
color(rgb_label ~= color_labels(count)) = 0;
segmented_images(:,:,:,count) = color;

%Choose 1 to show the color of the fiducial points
imshow(segmented_images(:,:,:,1));
end

By the way, should I create another question about the angle problem, separated from this one? (As I see you are only replying to the problem about the color).

[Image Analyst]

Yes - that was a copy/paste issue that I forgot to change.

Answer 3

[HecLondon]

Hi Image Analyst,

Many thanks for your answer. I think I didn't explain myself properly: I would like to get the coordinates and the colour value of a pixel by clicking on the mouse directly into the workspace, without having to copy. In other words: I want to do automatic colorsegmentation, so I want the user to just click in a fiducial point, save this colour automatically in workspace and then use it somewhere else in the script. Is this possible?

[Image Analyst]

Yes, see my color segmentation tutorial: http://www.mathworks.com/matlabcentral/fileexchange/31118-color-segmentation-by-delta-e-color-difference You manually pick a color (draw some regions) and then it finds all colors like that in the picture.

Variables in MATLAB are always "saved" in a workspace - the workspace that is in scope, which usually means the function workspace if you're in a function or the base workspace if you're in a script. There is no need to save anything to the base workspace from a function, though there are functions to do that. This is almost always recommended against. Why do you think you have a special need to do that? In many years of writing hundreds of complex MATLAB programs of thousands of lines I've never needed to do that.

[HecLondon]

Ok, let me explain properly what I'm trying to do:

I want to get just by clicking in a fiducial point in my image the location and the color information. I can do this already with impixel, and this information is in workspace so I can use it later in the script (is this something silly about workspace? Do I have a lack of understanding here? I just want to use these values in the code to calculate other things, all automatic, the user having to click the fiducial point only).

The two challenges then are:

-How can I get, with impixel or something else, the values a and b of the Lab colorspace instead of the rgb values (I changed my image to Lab colorspace and use impixel but I stil get the rgb values...). My idea is to find the formula to transform rgb to a,b and do this in the code, do you have any other suggestion for this?

-Regarding the location of the pixel: I want to label and get the centroids of the fiducial points in a clockwise direction, the starting point being the one the user clicked at the beginning. In order to make this the 0 angle, I need to change the atan2 function or find something else, which allows me to rotate to axis to place the 0 angle where I want, is it possible to do this?

[Image Analyst]

I'd use ginput(1) to have the user select a single point. impixel() is kind of a pain for that. Then get the row and column from that, something like this untested code off the top of my head:

[x,y] = ginput(1);
row = int32(y);
column = int32(x);
labColor = labImage(row, column, :);

[HecLondon]

Thanks, but it is possible to get the L,a,b values instead of the rgb ones with labColor? I keep getting the rgb even I change my image to Lab (see below):

I = imread('skin2.JPG');  %read the image in I
RGB = imresize(I, 0.2);
cform = makecform('srgb2lab');
lab_fabric = applycform(RGB,cform);
[x,y] = ginput(1);
row = int32(y);
column = int32(x);
labColor = lab_fabric(row, column, :);

Also, following your comment above about the angles. Sorry for the misunderstanding: the angle problem is not related with the colors. Once I have the color of the pixel and I do color segmentation, I end up with a black and white image of the fiducial points. Now, I want to put labels in these points and get the centroids in a clockwise direction, starting with the point clicked at the beginning (starting angle).

The only idea I had to order the points in a clockwise direction was using the angles of the points, and order them in a increasing order. The problem is that I don't know how to change the atan2 function (or use another one), to put the zero angle where the first fiducail point is (it seems the axis are always horizontal). Do you have any idea to go around this problem?

[Image Analyst]

Let's continue the conversation under my answer, not yours. But it looks like that should give you the lab values, not the rgb values because lab_fabric is a different array than RGB.