How to count the average rgb value of label region from original image?

I have a label matrix image and label 1 is the border

    labelimg =                  original image rgb=

    1 2 3 4 5 6                     1             2             3             4             5             6             
  1 3 3 3 3 3 3             1 (111,111,111) (111,111,111) (111,111,111) (111,111,111) (111,111,111) (112,112,112) 
  2 1 1 1 1 1 1             2 (000,000,000) (000,000,000) (000,000,000) (000,000,000) (000,000,000) (000,000,000)
  3 1 2 2 1 4 1             3 (000,000,000) (222,222,222) (222,222,222) (000,000,000) (223,221,223) (000,000,000)
  4 1 1 1 1 4 1             4 (000,000,000) (000,000,000) (000,000,000) (000,000,000) (223,221,223) (000,000,000)
  5 5 5 5 1 1 1             5 (221,221,221) (223,223,223) (223,223,223) (000,000,000) (000,000,000) (000,000,000)
  6 5 5 5 5 5 5             6 (221,221,221) (221,221,221) (221,221,221) (221,221,221) (221,221,221) (221,221,221)

I want count the average(median is more complicated,I want try on average first see whethere can solve my problem or not) rgb value of label region from original image~how to do this?

everytime I loop horizontal through the labelimage, if read 3 then count of 3= +1

I need get the count table so I can count

average=all pixel value of labelx/count of label x

average pixel value of label x=( r,g,b )

count = 
label    count
2          2   
3          6
4          2
5          9

label region  average rgb(use the round to integer)
 2            (111,111,111)
 3            (r,g,b)
 4            (r,g,b)
 5            (r,g,b)

label is the label image what i done and elephant is the original image

Answer 1

[Image Analyst]

First you need to split up your image into individual color channels:

% Extract the individual red, green, and blue color channels.
redChannel = rgbImage(:, :, 1);
greenChannel = rgbImage(:, :, 2);
blueChannel = rgbImage(:, :, 3);

Then you need to get a binary image of each label and then you can get the means for that particular label number.

for yourNumber = 1 : max(labeledImage(:))
    thisLabel = labeledImage == yourNumber; % for example 2 or 3 or whatever region you want.
    meanR(yourNumber) = mean(redChannel(thisLabel));
    meanG(yourNumber) = mean(greenChannel(thisLabel));
    meanB(yourNumber) = mean(blueChannel(thisLabel));
end

Again, like your other questions, this is not what you want to do, but it does what you asked.

[Tan Wen Kun]

It is the thing I want to do,after got mean then I want try merge the label together like label 1 and label 2 comparing rgb whether is +-5 or not.

If yes then I reassign label 2 to label 1.

After reassign all the label then I do the clearborder so I can get the lousy segmentation for my image. That my work done. I also regret get a image segmentation topic, it made my life worst because never touch matlab and image segmentation before, I only got 4week to learn it so this the reason I always the weird question you think.

I want turn label to final result is get almost like that but the label is already reassign so the object maybe only got few label color only

[Image Analyst]

It's really not what you want to do - you just don't realize it yet. It's a wild goose chase - a dead end. But sometimes by doing those things, that's how you learn.