Matrix representations of the multidimensional overlap and add technique

The authors present sparse matrix representations for multidimensional Agarwal-Burrus nesting. It is shown that the resulting sparse matrices can be expressed in terms of tensor product decompositions. The formulations derived provide a highly compact representation of this approach to nesting. The tensor product approach also provides the needed structure to easily derive several variant schemes that are well suited for use on many types of computer architectures.<
>