Comments on: K-Nearest Neigbors and Radius (Range) Search http://www.advancedmcode.org/k-nearest-neigbours-search.html Open Blog with: Engineering Solutions, Algorithms, Advanced Matlab Source Code and Science related contents Thu, 11 Mar 2010 12:10:59 +0100 http://wordpress.org/?v=2.9.1 hourly 1 By: Luigi Giaccari http://www.advancedmcode.org/k-nearest-neigbours-search.html/comment-page-1#comment-304 Luigi Giaccari Mon, 12 Oct 2009 17:53:49 +0000 http://www.advancedmcode.org/?p=241#comment-304 Try this: <pre lang="MATLAB"> N=1000;%number of reference points Nq=10;%number of query points dim=3;%dimension of points r=.2;%Search radius p=rand(N,dim); qp=rand(Nq,dim); A=sparse(N,N);%building the sparse matrix for i=1:Nq idc=BruteSearchMex(p',qp(i,:)','r',r);%%Radius Neighbor m=length(idc);%number of neighbours found A([i*ones(1,m),idc],[idc,i*ones(1,m)])=1;%add element to matrix end </pre> Consider that sparse matrix in Maltab is class defined for doubles. So if you want a logical amtrix you need to use: <pre lang="MATLAB"> A=false(N,N); </pre> It depends on what are your requirements. Second, yes radius can be different. Radisu search supports only one query time so you have to call the rourtine each loop. Good luck Luigi Try this:

N=1000;%number of reference points
Nq=10;%number of query points
dim=3;%dimension of points
r=.2;%Search radius
 
p=rand(N,dim);
qp=rand(Nq,dim);
 
A=sparse(N,N);%building the sparse matrix
 
for i=1:Nq
idc=BruteSearchMex(p',qp(i,:)','r',r);%%Radius Neighbor
m=length(idc);%number of neighbours found
A([i*ones(1,m),idc],[idc,i*ones(1,m)])=1;%add element to matrix
end

Consider that sparse matrix in Maltab is class defined for doubles. So if you want a logical amtrix you need to use:

A=false(N,N);

It depends on what are your requirements.

Second, yes radius can be different. Radisu search supports only one query time so you have to call the rourtine each loop.

Good luck

Luigi

]]> By: Arne Bøckmann http://www.advancedmcode.org/k-nearest-neigbours-search.html/comment-page-1#comment-302 Arne Bøckmann Mon, 12 Oct 2009 16:29:16 +0000 http://www.advancedmcode.org/?p=241#comment-302 Could you provide an example on how to find the sparse connectivity matrix with a specified cut-off radius ? Could this radius also differ from point to point? Could you provide an example on how to find the sparse connectivity matrix with a specified cut-off radius ? Could this radius also differ from point to point?

]]>