shift dim matlab,mean shift图像分割的matlab代码，求解释啊，看不懂

该楼层疑似违规已被系统折叠 隐藏此楼查看此楼

function B1 = smooth(A,hr,hs)

[row,col,dim] = size(A);

A = double(A);

A1 = A(:,:,1);

A2 = A(:,:,2);

A3 = A(:,:,3);

B = A;

h = 2*hs;%Gaussian Kernel

u = 2*h+1;

kernel = ones(u,u);

for i=1:1:u

for j=1:1:u

kernel(i,j)=(i-h-1)^2+(j-h-1)^2;

end

end

kernel = exp(-kernel./(hs*hs));%Smooth the original graph

for i = 1:row

for j = 1:col

x1 = max([1,h+2-i]);

y1 = max([1,h+2-j]);

x2 = min([2*h+1,row+h+1-i]);

y2 = min([2*h+1,col+h+1-j]);

Partkernel = kernel(x1:x2,y1:y2);

x3 = i-h-1+x1; %The codes above seem to be very complicated, but note that we need

y3 = j-h-1+y1; %to check the boundary of the graph. I will explain this in the project

x4 = i+x2-h-1; %report more clearly. These codes prevent such a condition that the

y4 = j+y2-h-1; %pixels which will be operating in the kernel exceeds the boundary of the picture.

for t=1:1:5 %Every time, the vector moves towards the destination a litter. Here, the iteration is 5 times. It is suffcient to converge.

Part = A1(x3:x4,y3:y4);

Difference = Part - A1(i,j);

Difference = Difference.^2;

R = exp(-Difference./(hr.^2));

weight = R.*Partkernel;

value = sum(sum(Part.*weight));

sumnumber = sum(sum(weight));

A1(i,j) = value/sumnumber; Part = A2(x3:x4,y3:y4);

Difference = Part - A2(i,j);

Difference = Difference.^2;

R = exp(-Difference./(hr.^2));

weight = R.*Partkernel;

value = sum(sum(Part.*weight));

sumnumber = sum(sum(weight));

A2(i,j) = value/sumnumber; Part = A3(x3:x4,y3:y4);

Difference = Part - A3(i,j);

Difference = Difference.^2;

R = exp(-Difference./(hr.^2));

weight = R.*Partkernel;

value = sum(sum(Part.*weight));

sumnumber = sum(sum(weight));

A3(i,j) = value/sumnumber;

end

B(i,j,1)=A1(i,j);

B(i,j,2)=A2(i,j);

B(i,j,3)=A3(i,j);

end

end

B1 = uint8(B);