A convex relaxation approach for computing minimal partitions

By Thomas Pock, Antonin Chambolle, Daniel Cremers, Horst Bischof. 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009

303 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/96b28d117fb344109eb52984ecbc5e61e166eacf

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Beyond pixels: A comprehensive survey from bottom-up to semantic image segmentation and cosegmentation (2015, methodology): “By assuming the region is piecewise constant, the model is further simpliﬁed to the continuous Potts model, which has convexiﬁed solvers [53], [54].”
Joint motion estimation and segmentation of complex scenes with label costs and occlusion modeling (2012, methodology): “Results are shown on several classical motion segmentation and optical flow examples.”
Learning full pairwise affinities for spectral segmentation (2010, background): “Recently, the main research directions are focused on deterministic annealing [21], stochastic clustering [13][26], mode-seeking methods like mean shift [5][4], mixture model [23][24], rate distortion [22], graph theoretic approaches [27][9][33], and other variational methods [20][8].”
Normalized Cut Loss for Weakly-Supervised CNN Segmentation (2018, background): “Due to importance of lowlevel regularizers for segmentation and computer vision in general, there are many choices available in the literature [19, 11, 9, 43, 23, 34, 39, 30, 15, 46].”
A Survey of Optimization Methods From a Machine Learning Perspective (2019, methodology): “In the field of computer vision, a convex relaxation method was proposed to calculate minimal partitions [136].”
Dense Semantic 3D Reconstruction (2017, background): “With this notation in place, the convex relaxation, originating from [5], [30], of the labeling problem in a continuous volumetric domainV reads as”
On Regularized Losses for Weakly-supervised CNN Segmentation (2018, methodology): “Exploring other relaxations of CRF as losses [18,19,20,21,22,23] and corresponding eﬃcient gradient computation is left for future work.”
Minimizing Energies with Hierarchical Costs (2012, background): “A number of variational formulations of E have recently been developed with continuous analogues of D + V (e.g. Pock et al. 2008, 2009; Olsson et al. 2009) and of D + V + H (Yuan and Boykov 2010).”
Simultaneous Optical Flow and Intensity Estimation from an Event Camera (2016, methodology): “Variational methods have been very successful to include smoothness priors such as TV-L1 [18], which approximates natural image statistics [20], while leaving the optimisation problem convex.”
Large-Scale Semantic 3D Reconstruction: An Adaptive Multi-resolution Model for Multi-class Volumetric Labeling (2016, background): “One limitation of that model is its huge memory consumption, which we address with our spatially adaptive scheme, without loss in quality.”