A p-Laplacian Inspired Method for Graph Cut

One of the approaches to optimally divide the cellular networks of the next generation, with a large number of base-stations, is based on the graph cuts theory. Graph cut aims at identifying and arranging the vertices into non-overlapping subsets. It is expected that data in each subset share some relative similarities. This paper reviews the graph clustering approach based on the minimum cut with the p-Laplacian, using a simplified method for Fiedler vector thresholding. The proposed method is inspired by the developments within the compressive sensing theory.