Bottleneck Potentials in Markov Random Fields

@article{Abbas2019BottleneckPI,
  title={Bottleneck Potentials in Markov Random Fields},
  author={Ahmed Abbas and Paul Swoboda},
  journal={2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2019},
  pages={3174-3183}
}
  • Ahmed Abbas, P. Swoboda
  • Published 17 April 2019
  • Computer Science
  • 2019 IEEE/CVF International Conference on Computer Vision (ICCV)
We consider general discrete Markov Random Fields(MRFs) with additional bottleneck potentials which penalize the maximum (instead of the sum) over local potential value taken by the MRF-assignment. [] Key Result We empirically show efficacy of our approach on large scale seismic horizon tracking problems.
Two Polynomial Time Graph Labeling Algorithms Optimizing Max-Norm-Based Objective Functions
TLDR
This work considers a problem of finding an optimal binary labeling for the objective function defined as the max-norm over a set of local costs of a form that naturally appears in image processing, and proposes two algorithms that find such optimal labelings.
Structured Prediction Problem Archive
TLDR
This work collects in one place a large number of datasets in easy to read formats for a diverse set of problem classes to facilitate algorithm development for their numerical solution.

References

SHOWING 1-10 OF 52 REFERENCES
A Linear Programming Approach to Max-Sum Problem: A Review
  • Tomáš Werner
  • Computer Science
    IEEE Transactions on Pattern Analysis and Machine Intelligence
  • 2007
TLDR
This work reviews a not widely known approach to the max-sum labeling problem, developed by Ukrainian researchers Schlesinger et al. in 1976, and shows how it contributes to recent results, most importantly, those on the convex combination of trees and tree-reweighted max-product.
Energy minimization for linear envelope MRFs
  • Pushmeet Kohli, M. P. Kumar
  • Computer Science
    2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition
  • 2010
TLDR
It is shown how the minimization of energy functions with upper envelope potentials leads to a difficult minmax problem, and a new message passing algorithm is proposed that solves a linear programming relaxation of the problem.
Lagrangian-based methods for finding MAP solutions for MRF models
TLDR
This paper explores the use of Lagrange relaxation (LR) methods for solving the maximum a posteriori (MAP) solutions from noisy images based on a prior Markov random field (MRF) model as an integer linear programming (ILP) problem.
Min-Max Propagation
TLDR
It is shown that for “any” high-order function that can be minimized in O(ω), the min-max message update can be obtained using an efficient O(K(ω + log(K)) procedure, where K is the number of variables.
Power Watershed: A Unifying Graph-Based Optimization Framework
TLDR
This work extends a common framework for graph-based image segmentation that includes the graph cuts, random walker, and shortest path optimization algorithms and proposes a new family of segmentation algorithms that fixes p to produce an optimal spanning forest but varies the power q beyond the usual watershed algorithm.
A Comparative Study of Modern Inference Techniques for Structured Discrete Energy Minimization Problems
TLDR
An empirical comparison of more than 27 state-of-the-art optimization techniques on a corpus of 2453 energy minimization instances from diverse applications in computer vision suggests that polyhedral methods and integer programming solvers are competitive in terms of runtime and solution quality over a large range of model types.
MRF Energy Minimization and Beyond via Dual Decomposition
TLDR
It is shown that by appropriately choosing what subproblems to use, one can design novel and very powerful MRF optimization algorithms, which are able to derive algorithms that generalize and extend state-of-the-art message-passing methods, and take full advantage of the special structure that may exist in particular MRFs.
Network Flow Algorithms for Structured Sparsity
TLDR
This work considers a class of learning problems that involve a structured sparsity-inducing norm defined as the sum of l∞-norms over groups of variables, and proposes an efficient procedure which computes its solution exactly in polynomial time.
Unified Framework for Semiring-based Arc Consistency and Relaxation Labeling
TLDR
Constraint Satisfaction Problem (CSP), including its soft modifications, is ubiquitous in artificial intelligence and related fields and a fundamental concept to tackle the CSP, as well as the SCSPs with idempotent semiring multiplication, are arc consistency algorithms, also known as relaxation labeling.
Combining the Shortest Paths and the Bottleneck Paths Problems
TLDR
This paper solves the Single Source Shortest Paths for All Flows (SSSP-AF) problem on directed graphs with unit edge costs in O(mn) worst case time bound and presents two algorithms to solve SSSP-af on directed graph with integer edge costs bounded by c in O (m2 + nc) and O( m2 + mn log (c/m) time bounds.
...
...