Blue-Noise Sampling of Signals on Graphs

In this work we introduce the concept of blue noise sampling, traditionally used in imaging applications, for bandlimited signals on graphs. We show how the spectral and vertex domain characterization of these patterns is connected with results about the quality of the sampling sets already existing in the literature. We provide numerical evidence that shows that these patterns are also competitive with respect to the state of the art sampling techniques in terms of the reconstruction error.