This is where navigation should be.


GSP_GRAD - Graph gradient

Usage

gr = gsp_grad(G,s)

Input parameters

G Graph structure
s Signal living on the nodes

Output parameters

gr Gradient living on the edges

Description

For the non normalized Laplacian, the gradient of the node signal \(f\) evaluated at the edge linking \(x\) and \(y\) is given by:

\begin{equation*} \nabla f(x,y) = \sqrt{ w(x,y) } \left( f(x) - f(y) \right) \end{equation*}

Before using this function, you need to call the function:

G = gsp_adj2vec(G)