基于 K-means 聚类算法的图像区域分割.zip

function [center]=searchcenter(X,kratio) [n,~]=size(X); isleft=true(n,1); count=zeros(n,1); center=[]; kind=0; dist=0; for i=1:n for j=i+1:n dist=dist+weightdist(X(i,:),X(j,:)); end end dist=dist/((n-1)*(n-1)/2); radius1=dist*kratio(1); radius2=dist*kratio(2); while any(isleft) for i=1:n count(i)=0; if isleft(i) for j=1:n if isleft(j) dist=weightdist(X(i,:),X(j,:)); count(i)=count(i) + dist<=radius1; end end end end [~,locs]=max(count); iscenter=true; for i=1:kind dist=weightdist(X(locs,:),center(i,:)); iscenter=iscenter && dist>=radius2; if ~iscenter break; end end if iscenter kind=kind+1; center(end+1,:)=X(locs,:); for i=1:n if isleft(i) dist=weightdist(X(i,:),X(locs,:)); if dist <= radius1 isleft(i)=false; end end end else isleft(locs)=false; end end