Skip to main content


Fig. 4 | EURASIP Journal on Advances in Signal Processing

Fig. 4

From: Stationary time-vertex signal processing

Fig. 4

Estimation error \({\tilde {\mathbf {E}}\left [{{\left \|\ddot {\mathbf {H}} - \mathbf {H}\right \|}_{F}} \right ]}/{{\left \|\mathbf {H}\right \|}_{F}}\) as a function of the number of realizations and number of vertices. Even an approximate knowledge of the graph enables us to make good estimates of the covariance (and PSD) from few realizations. The joint stationarity prior becomes especially meaningful when the number of variables (N,T) increases. The benefit also holds for a noisy graph (SNR = 10dB). a N=10, T=10b N=100, T=10

Back to article page