A Discriminative View of MRF Pre-processing Algorithms

@article{Wang2017ADV,
  title={A Discriminative View of MRF Pre-processing Algorithms},
  author={Chen Wang and Charles Herrmann and R. Zabih},
  journal={2017 IEEE International Conference on Computer Vision (ICCV)},
  year={2017},
  pages={5505-5514}
}
  • Chen Wang, Charles Herrmann, R. Zabih
  • Published 2017
  • Computer Science
  • 2017 IEEE International Conference on Computer Vision (ICCV)
  • While Markov Random Fields (MRFs) are widely used in computer vision, they present a quite challenging inference problem. MRF inference can be accelerated by preprocessing techniques like Dead End Elimination (DEE) [8] or QPBO-based approaches [18, 24, 25] which compute the optimal labeling of a subset of variables. These techniques are guaranteed to never wrongly label a variable but they often leave a large number of variables unlabeled. We address this shortcoming by interpreting pre… CONTINUE READING

    References

    SHOWING 1-10 OF 53 REFERENCES
    Relaxation-Based Preprocessing Techniques for Markov Random Field Inference
    • Chen Wang, R. Zabih
    • Mathematics, Computer Science
    • 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
    • 2016
    • 4
    • PDF
    Optimizing Binary MRFs via Extended Roof Duality
    • 490
    • PDF
    On partial optimality in multi-label MRFs
    • 72
    • Highly Influential
    • PDF
    Dynamic Hybrid Algorithms for MAP Inference in Discrete MRFs
    • 44
    • PDF
    Alphabet SOUP: A framework for approximate energy minimization
    • 28
    • PDF
    Continuous Multiclass Labeling Approaches and Algorithms
    • 123
    • PDF
    A Comparative Study of Energy Minimization Methods for Markov Random Fields
    • 482
    • PDF
    A Comparative Study of Modern Inference Techniques for Discrete Energy Minimization Problems
    • 207
    • PDF
    Fusion Moves for Markov Random Field Optimization
    • 232
    • PDF
    A bundle approach to efficient MAP-inference by Lagrangian relaxation
    • 53
    • PDF