[Insight-users] sparse matrices

[Rupert Brooks]

Hi folks,

I've been wondering if theres a proper way to use sparse matrix
approaches for numerical calculations in ITK.  To my way of thinking,
a sparse matrix should be a memory-efficient subclass of itk::Array2D,
which would overload the operators accordingly so that matrix/vector
multiplication, etc would be efficient when possible due to sparsity.

There is a vnl_sparse_matrix, but it doesnt seem to be a subclass of
vnl_matrix, or to have overloaded matrix multiplication operators.

Ideally, i'd like something thats polymorphic with a vnl_matrix, and
or an itk::Array2D.  I'd definitely consider writing one, but first...
have i missed a better way?   Has someone already done it?

Cheers,
Rupert B.
-- 
--------------------------------------------------------------
Rupert Brooks
McGill Centre for Intelligent Machines (www.cim.mcgill.ca)
Ph.D Student, Electrical and Computer Engineering
http://www.cyberus.ca/~rbrooks