Parallel and Distributed Vision Algorithms Using Dual Decomposition

Summary, in English

We investigate dual decomposition approaches for optimization problems arising in low-level vision. Dual decomposition can be used to parallelize existing algorithms, reduce memory requirements and to obtain approximate solutions of hard problems. An extensive set of experiments are performed for a variety of application problems including graph cut segmentation, curvature regularization and more generally the optimization of MRFs. We demonstrate that the technique can be useful for desktop computers, graphical processing units and supercomputer clusters. To facilitate further research, an implementation of the decomposition methods is made publicly available.

• Författare: Petter Strandmark; Fredrik Kahl; Thomas Schoenemann
• Avdelning/ar: Matematik LTH; Mathematical Imaging Group; ELLIIT: the Linköping-Lund initiative on IT and mobile communication
• Publiceringsår: 2011
• Språk: Engelska
• Sidor: 1721-1732
• Publikation/Tidskrift/Serie: Computer Vision and Image Understanding
• Volym: 115
• Issue: 12
• Länkar: http://www.sciencedirect.com/science/article/pii/S1077314211001652; http://dx.doi.org/10.1016/j.cviu.2011.06.012
• Dokumenttyp: Artikel i tidskrift
• Förlag: Elsevier
• Ämne: Mathematics; Computer Vision and Robotics (Autonomous Systems)
• Nyckelord: Graph cuts; Dual decomposition; Parallel; MRF; MPI; GPU
• Status: Published
• Forskningsgrupp: Mathematical Imaging Group
• ISBN/ISSN/Övrigt: ISSN: 1077-3142