I am trying to list down applications where having a rank-1 matrix is advantageous. I know only of 2D convolution which boils down to a series of 1D convolutions if filter response is separable.
Can members add any other application here?
I am trying to list down applications where having a rank-1 matrix is advantageous. I know only of 2D convolution which boils down to a series of 1D convolutions if filter response is separable.
Can members add any other application here?
Copyright © 2021 JogjaFile Inc.
The Spectral Theorem is about decomposing a matrix as a linear combination of rank-one orthogonal projections.
Diagonalization is about decomposing a matrix as a linear combination of rank-one idempotents.