The recently introduced random walker segmentation algorithm by Grady and Funka-Lea (2004) has been shown to have desirable theoretical properties and to perform well on a wide variety of images in practice. However, this algorithm requires user-specified labels and produces a segmentation where each segment is connected to a labeled pixel. We show that incorporation of a nonparametric probability density model allows for an extended random walkers algorithm that can locate disconnected objects and does not require user-specified labels. Finally, we show that this formulation leads to a deep connection with the popular graph cuts method by Boykov et al. (2001) and Wu and Leahy (1993).
Multilabel random walker image segmentation using prior models
2005-01-01
360852 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Multilabel partition moves for MRF optimization
British Library Online Contents | 2013
|Joint Prior Models of Neighboring Objects for 3D Image Segmentation
British Library Conference Proceedings | 2004
|Robust and efficient image segmentation approaches using Markov random field models
British Library Online Contents | 2003
|