Continuous multiclass labeling approaches and algorithms
From MaRDI portal
total variationconvex relaxationsegmentationnonsmooth optimizationvariational methodssaddle point problemsplitting methodscontinuous cut
Numerical optimization and variational techniques (65K10) Convex programming (90C25) Combinatorial optimization (90C27) Computing methodologies for image processing (68U10) Numerical aspects of computer graphics, image analysis, and computational geometry (65D18) Numerical methods of relaxation type (49M20)
Abstract: We study convex relaxations of the image labeling problem on a continuous domain with regularizers based on metric interaction potentials. The generic framework ensures existence of minimizers and covers a wide range of relaxations of the originally combinatorial problem. We focus on two specific relaxations that differ in flexibility and simplicity -- one can be used to tightly relax any metric interaction potential, while the other one only covers Euclidean metrics but requires less computational effort. For solving the nonsmooth discretized problem, we propose a globally convergent Douglas-Rachford scheme, and show that a sequence of dual iterates can be recovered in order to provide a posteriori optimality bounds. In a quantitative comparison to two other first-order methods, the approach shows competitive performance on synthetical and real-world images. By combining the method with an improved binarization technique for nonstandard potentials, we were able to routinely recover discrete solutions within 1%--5% of the global optimum for the combinatorial image labeling problem.
Recommendations
- Tight convex relaxations for vector-valued labeling
- scientific article; zbMATH DE number 6025169
- A survey and comparison of discrete and continuous multi-label optimization approaches for the Potts model
- Global minimization for continuous multiphase partitioning problems using a dual approach
- Optimality bounds for a variational relaxation of the image partitioning problem
Cited in
(44)- Assignment flows
- Optimality bounds for a variational relaxation of the image partitioning problem
- Shearlets: theory and applications
- Iterative multiplicative filters for data labeling
- Variational image segmentation models involving non-smooth data-fidelity terms
- Sequential convex programming for computing information-theoretic minimal partitions: nonconvex nonsmooth optimization
- scientific article; zbMATH DE number 1844608 (Why is no real title available?)
- Image labeling by assignment
- Optimization of robust loss functions for weakly-labeled image taxonomies
- Inverse scale space iterations for non-convex variational problems: the continuous and discrete case
- Bregman-proximal augmented Lagrangian approach to multiphase image segmentation
- Convex relaxation of a class of vertex penalizing functionals
- Piecewise-smooth image segmentation models with \(L^1\) data-fidelity terms
- Convex multiclass segmentation with shearlet regularization
- Linkage between piecewise constant Mumford-Shah model and Rudin-Osher-Fatemi model and its virtue in image segmentation
- An efficient and versatile variational method for high-dimensional data classification
- An overview of SaT segmentation methodology and its applications in image processing
- Effective two-stage image segmentation: a new non-Lipschitz decomposition approach with convergent algorithm
- Image labeling based on graphical models using Wasserstein messages and geometric assignment
- Completely convex formulation of the Chan-Vese image segmentation model
- Image segmentation based on the hybrid total variation model and the K-means clustering strategy
- Multi-class transductive learning based on \(\ell^1\) relaxations of Cheeger cut and Mumford-Shah-Potts model
- Global minimization for continuous multiphase partitioning problems using a dual approach
- Iterative Potts minimization for the recovery of signals with discontinuities from indirect measurements: the multivariate case
- Multicuts and perturb \& MAP for probabilistic graph clustering
- Lifting methods for manifold-valued variational problems
- On the coupled continuous knapsack problems: projection onto the volume constrained Gibbs \(N\)-simplex
- Modelling convex shape priors and matching based on the Gromov-Wasserstein distance
- Consistent procedures for multiclass classification of discrete diffusion paths
- Midrange geometric interactions for semantic segmentation. Constraints for continuous multi-label optimization
- Globally optimal joint image segmentation and shape matching based on Wasserstein modes
- A narrow band method for the convex formulation of discrete multilabel problems
- Multilabel classifiers with a probabilistic thresholding strategy
- Coupling image restoration and segmentation: a generalized linear model/Bregman perspective
- An introduction to continuous optimization for imaging
- Continuous-domain assignment flows
- A spatially continuous max-flow and min-cut framework for binary labeling problems
- Recent advances of variational models in medical imaging and applications to computer aided surgery
- Directed acyclic graph continuous max-flow image segmentation for unconstrained label orderings
- Convex Regularization of Discrete-Valued Inverse Problems
- Tight convex relaxations for vector-valued labeling
- Graphmax for text generation
- Discrete and continuous models for partitioning problems
- Probabilistic correlation clustering and image partitioning using perturbed multicuts
This page was built for publication: Continuous multiclass labeling approaches and algorithms
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3113848)
