Problem of large adjacency matrix

Hi,

I need to deal with a very large adjacency matrix in Matlab, i have a image matrix of size 321x481, and i want te generate the adjacency matrix from this image, its size is 321*481 x 321*481, but matlab give the message error Out of memory, can anybody give me a solution o that. My code is like that:

I= imread(Myimage);

A= Adjaceny_Matrix(I);

My treatement

Thank a Lot.

Answer 1

[John D'Errico]

Just wanting to solve a large problem is not enough.

You lack sufficient memory. What is the size of that matrix?

321*481 * 321*481*8
ans =
   1.9072e+11

So roughly 190 GIGABYTES of memory. Be careful not to make a copy of that matrix along the way too.

Computers are not infinitely large or infinitely fast.

[Mourchid]

So? there is a solution for that ??

Answer 2

[Walter Roberson]

You can use a sparse() array. Each entry is connected to 8 others if you are using standard 8-connectivity and you should be able to fit that easily.

However I recommend you reconsider whether you really need the adjacency matrix. For any one entry, the connected elements can be calculated fairly easily: for logical index K on the interior of the matrix, the logical index of the entry "above" is K-1, below is K+1, left is K-321, right is K+321, and diagonals are K-322, K+320, K-320, K+322 . With the connections easily calculated it does not seem appropriate to use a lot of memory to store them.

[Mourchid]

Thank you Wlater Roberson for your respense, can you please give me the matlab code to doing it. for example i have the matlab code:

I=imread(Myimage); what i will do after ?