SLIC算法的matlab代码

以下是 SLIC 算法的 MATLAB 代码示例。此代码旨在提供一个基本的 SLIC 实现。这个例子在图片的颜色空间上执行聚类。如果您想用其他空间做聚类（例如CIELAB），需要自行更改。在这个例子中，SLIC 算法为每个超像素分配一个标签，并计算每个超像素的位置和平均颜色值。

% Slice superpixel clustering
clear all; close all; clc;

% Load input image
inputImg = imread('input.jpg');
[nRows,nCols,~] = size(inputImg);

% Parameters to set
k = 600;                    % Desired number of superpixels
m = 30;                 % Compactness factor. Higher -> more compact superpixels.
N = nRows * nCols;          % Total number of pixels

% Compute initial positions of cluster centers
S = floor(sqrt(N/k));
nx = floor(nCols/S);
ny = floor(nRows/S);
Cx = zeros(nx*ny,1);
Cy = zeros(nx*ny,1);
cx = S/2:S:nCols-S/2;
cy = S/2:S:nRows-S/2;
[X,Y] = meshgrid(cx,cy);
Cx = reshape(X,[],1);
Cy = reshape(Y,[],1);

% Initialize cluster assignments and distances
assignments = zeros(N, 1);
distances = inf(N, 1);

% Convert input RGB image to L*a*b* color space
inputImg = double(inputImg);
inputImg = inputImg / max(inputImg(:));
inputImgLab = RGB2Lab(inputImg);

% Iterative optimization
for iter = 1:10
    distances(:) = inf;

    for c = 1:size(Cx,1)
        cx = round(Cx(c));
        cy = round(Cy(c));
        
        % Define search window
        xmin = max(cx-S,1);
        ymin = max(cy-S,1);
        xmax = min(cx+S,nCols);
        ymax = min(cy+S,nRows);      
        window = inputImgLab(ymin:ymax,xmin:xmax,:);
 
        % Compute distances to cluster center
        dx = (window(:,:,2)-inputImgLab(cy,cx,2)).^2;
        dy = (window(:,:,3)-inputImgLab(cy,cx,3)).^2;
        dlab = dx + dy./m^2;
        idx = sub2ind([ymax-ymin+1,xmax-xmin+1],repmat(cy-ymin+1,[1,xmax-xmin+1]),repmat(cx-xmin+1,[1,ymax-ymin+1]),1:(xmax-xmin+1)*(ymax-ymin+1));
        
        % Assign pixels to nearest cluster center
        mask = dlab < distances(idx);
        distances(idx(mask)) = dlab(mask);
        assignments(idx(mask)) = c;
    end
    
    % Update cluster centers
    for c = 1:size(Cx,1)
        idx = find(assignments==c);
        [y,x] = ind2sub([nRows,nCols],idx);
        Cx(c) = sum(x) ./ length(idx);
        Cy(c) = sum(y) ./ length(idx);
        region = inputImgLab(min(y):max(y),min(x):max(x),:);
        C(c,1) = mean(region(:,:,1),'all');
        C(c,2) = mean(region(:,:,2),'all');
        C(c,3) = mean(region(:,:,3),'all');
    end
end

% Display segmented image
outputImg = zeros(nRows,nCols,3);
[~,~,idx] = unique(assignments);
for i = 1:N
    outputImg(i) = C(idx(i),1);
    outputImg(i) = C(idx(i),2);
    outputImg(i) = C(idx(i),3);
end
figure,imshow(outputImg);
title('Segmented image');

% Convert RGB image to L*a*b* color space
function img_out = RGB2Lab(img_in)
R = img_in(:,:,1);
G = img_in(:,:,2);
B = img_in(:,:,3);

% Convert RGB to XYZ
X = 0.412453*R + 0.357580*G + 0.180423*B;
Y = 0.212671*R + 0.715160*G + 0.072169*B;
Z = 0.019334*R + 0.119193*G + 0.950227*B;

% Convert XYZ to L*a*b*
Xn = 0.950456;
Yn = 1.000000;
Zn = 1.088754;
L = 116*(Y/Yn).^(1/3) - 16;
a = 500*(X/Xn).^(1/3) - 500*(Y/Yn).^(1/3);
b = 200*(Y/Yn).^(1/3) - 200*(Z/Zn).^(1/3);

% Return Lab image
img_out(:,:,1) = L;
img_out(:,:,2) = a;
img_out(:,:,3) = b;
end

希望能帮到你！