Affiliations: Université de Caen, GREYC UMR-6072, Campus II,
Bd Marechal Juin BP 5186, 14032 Caen cedex, France.
[email protected]; [email protected]
Note: [] Address for correspondence: Université de Caen, GREYC
UMR-6072, Campus II, Bd Marechal Juin BP 5186, 14032 Caen cedex, France
Abstract: Hypergraphs are a large generalisation of graphs; they are now used
for many low-level image processing, by example for noise reduction, edge
detection and segmentation [3, 4, 7]. In this paper we define a generic 2D and
3D-image representation based on a hypergraph. We present the mathematical
definition of the hypergraph representation of an image and we show how this
representation conducts to an efficient lossless compression algorithm for 2D
and 3D-images. Then we introduce both 2D and 3D version of the algorithm and we
give some experimental results on some various sets of images: 2D photo, 2D
synthetic pictures, 3D medical images and some short animated sequences.
