Relaxed Cheeger Cut for Image Segmentation
Résumé
Motivated by recent advances in spectral clustering that show the relation between the non linear p-Laplacian graph operator and the Cheeger cut problem, we propose to study and apply this methodology for image segmentation. Based on a ℓ 1 relaxation of the initial clustering problem, we show that these methods can outperform usual well-known graph based approaches, e.g., min-cut/max-flow algorithm or ℓ 2 spectral cluster- ing, for unsupervised and very weakly supervised image segmentation. Experimental results demonstrate the benefits and the relevance of the proposed methodology especially for noisy image or when very few pixels are labeled for interactive image segmentation.
Domaines
Traitement des images [eess.IV]Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...