6533b85cfe1ef96bd12bcff3

RESEARCH PRODUCT

Neighborhood Hypergraph Partitioning for Image Segmentation

Soufiane RitalHocine CherifiSerge Miguet

subject

[ INFO ] Computer Science [cs]Computer Science::Computer Vision and Pattern RecognitionComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION[INFO]Computer Science [cs][INFO] Computer Science [cs]MathematicsofComputing_DISCRETEMATHEMATICS

description

International audience; The aim of this paper is to introduce a multilevel neighborhoodhypergraph partitioning for image segmentation. Our proposedapproach uses the image neighborhood hypergraph model introduced inour last works and the algorithm of multilevel hypergraphpartitioning introduced by George Karypis. To evaluate the algorithmperformance, experiments were carried out on a group of gray scaleimages. The results show that the proposed segmentation approachfind the region properly from images as compared to imagesegmentation algorithm using normalized cut criteria.Key words :Graph, Hypergraph, Neighborhood hypergraph, multilevel hypergraph partitioning, image segmentation, edge detection.

https://hal.archives-ouvertes.fr/hal-01589734