A non-local Chan-Vese model for sparse, tubular object segmentation

Anna Jezierska (1) , Olivia Miraucourt (2, 1) , Hugues Talbot (1) , Stéphanie Salmon (2) , Nicolas Passat (3)

1 LIGM - Laboratoire d'Informatique Gaspard-Monge
2 LMR - Laboratoire de Mathématiques de Reims
3 CRESTIC - Centre de Recherche en Sciences et Technologies de l'Information et de la Communication - EA 3804

To deal with the issue of tubular object segmentation, we propose a new model involving a non-local fitting term, in the Chan-Vese framework. This model aims at detecting objects whose intensities are not necessarily piecewise constant, or even composed of multiple piecewise constant regions. Our problem formulation exploits object sparsity in the image domain and a local ordering relationship between foreground and background. A continuous optimization scheme can then be efficiently considered in this context. This approach is validated on both synthetic and real retinal images. The non-local data fitting term is shown to be superior to the classical piecewise-constant model, robust to noise and to low contrast.

Mots clés: Index Terms— Variational image segmentation; non- local data fidelity; tubular structures; angiographic imaging

Domaines: Traitement des images [eess.IV]; Vision par ordinateur et reconnaissance de formes [cs.CV]; Imagerie médicale

Anna Jezierska, Olivia Miraucourt, Hugues Talbot, Stéphanie Salmon, Nicolas Passat. A non-local Chan-Vese model for sparse, tubular object segmentation. International Conference on Image Processing (ICIP), 2014, Paris, France. pp.907-911, ⟨10.1109/ICIP.2014.7025182⟩. ⟨hal-01695066⟩