Manifold-based mathematical morphology for graph signal editing of colored images and meshes

摘要: This paper presents a framework for morphological processing of graph signals and investigates its usage colored images meshes editing tasks. The proposed method enables, with the help construction manifold-based ordering color vectors, to define new representation in form an vectors index. relies on three steps: dictionary learning, manifold out sample extension. enables formulate operators graphs we demonstrate performance various image mesh applications (simplification, abstraction, enhancement).