K means Segmentation function

Question

Subhransu Sekhar Bhattacharjee 2021-4-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/813600-k-means-segmentation-function

回答： Sanju 2024-2-22

clc
img = imread('peppers.png'); 
T= rgb2lab(img);
T = im2uint8(T);
nrows = size(T,1);
ncols = size(T,2);
[X,Y] = meshgrid(1:ncols,1:nrows);
I = cat(3,T,X,Y);
K = 6;
P = imsegkmeans(T,K);
Z = labeloverlay(T,P);
L = imsegkmeans(I,K);
J = labeloverlay(T,L);
imshowpair(Z,J,'Montage');

I need to create a fucntion which replicates what imsegkmeans is doing right in this code. Could someone please explain with code.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Subhransu Sekhar Bhattacharjee 2021-4-26

@Image Analyst can you kindly help me with this?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Sanju 2024-2-22

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/813600-k-means-segmentation-function#answer_1414378

在 MATLAB Online 中打开

Hi Subhransu Sekhar Bhattacharjee,

The ‘imsegkmeans’ function in MATLAB performs k-means clustering on an image to segment it into ‘K’ distinct regions.

Here’s how you can create a function that replicates the functionality of ‘imsegkmeans’,

function P = myImsegkmeans(I, K)
    % Convert the input image to double precision
    I = im2double(I);
    % Reshape the image into a 2D matrix
    [nrows, ncols, ~] = size(I);
    X = reshape(I, nrows*ncols, []);
    % Perform k-means clustering
    [~, C] = kmeans(X, K);
    % Assign each pixel to the nearest cluster centroid
    [~, P] = pdist2(C, X, 'euclidean', 'Smallest', 1);
    P = reshape(P, nrows, ncols);
end

In this function, ‘I’ is the input image and ‘K’ is the desired number of clusters.

The function first converts the image to double precision using ‘im2double’. Then, it reshapes the image into a 2D matrix where each row represents a pixel and its color values. The ‘kmeans’ function is used to perform k-means clustering on the reshaped image, and the resulting cluster centroids are stored in ‘C’. Finally, each pixel in the image is assigned to the nearest cluster centroid using the ‘pdist2’ function, and the cluster assignments are reshaped back into the original image size.

You can use this function in your code by replacing the ‘imsegkmeans’ function call with ‘myImsegkmeans’,

P = myImsegkmeans(T, K);

You can also refer to the below documentation link if required,

https://www.mathworks.com/help/stats/kmeans.html

https://www.mathworks.com/help/matlab/ref/im2double.html

https://www.mathworks.com/help/matlab/ref/reshape.html

https://www.mathworks.com/help/stats/pdist2.html

Hope this Helps!