Find the Euclidean distance of each pixel from a user-selected pixel in color space and displays scaled images from 0 to the current max distance.

Hi,

Can someone help me? Sorry for bad english

I need to find the Euclidean distance of each pixel from a user-selected pixel in RGB space and other in Lab space. Then displays 2 scaled images either from 0 to the current max distance, or from 0 to the theoretically maximum possible distance in a given color space. The result should be like

I'm new to this

Here is my current code

I dont know if im on the right way

Thanks everyone

[KALYAN ACHARJYA]

I have read the description of the question twice, hope I have understood the question. I need confirmation before I answer the question.

Select any one coordinate (indices) within the image, get the ecludian distance to all pixel indices, right?

Is their any role of color space here, you may consider this irrespective of any color space?

What would be the result, size/format?

[KALYAN ACHARJYA]

Arsen Comments moved here

Hi, thanks for the reply

I need to calculate the Euclidean distance between pixel colors, not between their coordinates

With function "ginput(1)" i need to select one pixel and then find distance of each pixel from the user-selected pixel

first in RGB color space and then in LAB color space

Then I need to display 2 images:

one will display the Euclidean distance of each pixel from the user-selected pixel in RGB space and the other in space Lab.

Red dot is our selected pixel.

I need to scale the images with distances appropriately.

And the result should be like this in 2 different color spaces

Im really sorry for long reply

Best regards, Arsen

[KALYAN ACHARJYA]

"I need to calculate the Euclidean distance between pixel colors, not between their coordinates"

Distance measure between to positions/locations, the indices (row number and column number) of particular color level exactly represent the same.

[Arsen Driashkaba]

Sorry, im new to this. Can you help me please? Thanks a lot for replying

Answer 1

[Image Analyst]

Just do this

[x, y] = ginput(1);
row = round(y);
col = round(x);
rp = double(rgbImage(row, col, 1));
gp = double(rgbImage(row, col, 2));
bp = double(rgbImage(row, col, 3));
[r, g, b] = imsplit(double(rgbImage));
colorDistance = sqrt((r - rp).^2 + (g - gp).^2 + (b - bp).^2);
imshow(colorDistance, []);

See attached demo for Delta E demos. Delta E is the coor distance in LAB color space.

[Arsen Driashkaba]

Thank you for reply, it works for rgb

But i really dont understand how to do the same for Lab color space ;(

[Image Analyst]

Assuming you have your lab image in a variable called labImage, just make the straightforward changes:

[x, y] = ginput(1);
row = round(y);
col = round(x);
lp = double(labImage(row, col, 1));
ap = double(labImage(row, col, 2));
bp = double(labImage(row, col, 3));
[lChannel, aChannel, bChannel] = imsplit(double(rgbImage));
colorDistance = sqrt((lChannel - lp).^2 + (aChannel - ap).^2 + (bChannel - bp).^2);
imshow(colorDistance, []);

You might look up the deltae() function in MATLAB.

[Arsen Driashkaba]

I did the same just forgot [ ] as parameter of imshow(). Now it is working

Thank you very much for help!

[Arsen Driashkaba]

Hi sir,

please, can you help me with this one?

https://www.mathworks.com/matlabcentral/answers/725462-k-means-implementation-for-matrix-a-n-x-m-that-returns-a-vector-of-length-n-with-values-from-1-to-k

Thanks a lot