A noise cancellation algorithm based on hypergraph modeling

Abstract

Although the binary relations used in proximity graphs are relevant for many basic situations, they cannot represent the structuration process of digital images. In this paper we show that hypergraph theory is a more appropriate frame to describe the neighborhood relations that can be formalized between pixels. We illustrate the effectiveness of such a model by deriving a noise cancellation algorithm based on a basic combinatoric property of hypergraphs.