Complex Networks: A Networking and Signal Processing Perspective

Chapter 10Graph Signal Processing Approaches

The last chapter presented extension of classical signal processing operations such as convolution, translation, filtering, and modulation to the signals supported by complex networks. These operators were defined through graph Fourier transform (GFT). The GFT was defined by considering the eigenvectors of the (sym-metric) graph Laplacian matrix as the graph harmonics. In this approach to graph signal processing (GSP), the graph Laplacian played a fundamental role. However, GSP is not limited to the approach based on the graph Laplacian. There exist various approaches for GSP that define the GFT differently. For example, discrete signal processing on graphs (DSP_G) approach is equipped with the theory ...