Results 1 
4 of
4
Approximate labeling via graphcuts based on linear programming
 In Pattern Analysis and Machine Intelligence
, 2007
"... A new framework is presented for both understanding and developing graphcut based combinatorial algorithms suitable for the approximate optimization of a very wide class of MRFs that are frequently encountered in computer vision. The proposed framework utilizes tools from the duality theory of line ..."
Abstract

Cited by 74 (8 self)
 Add to MetaCart
(Show Context)
A new framework is presented for both understanding and developing graphcut based combinatorial algorithms suitable for the approximate optimization of a very wide class of MRFs that are frequently encountered in computer vision. The proposed framework utilizes tools from the duality theory of linear programming in order to provide an alternative and more general view of stateoftheart techniques like the αexpansion algorithm, which is included merely as a special case. Moreover, contrary to αexpansion, the derived algorithms generate solutions with guaranteed optimality properties for a much wider class of problems, e.g. even for MRFs with nonmetric potentials. In addition, they are capable of providing perinstance suboptimality bounds in all occasions, including discrete Markov Random Fields with an arbitrary potential function. These bounds prove to be very tight in practice (i.e. very close to 1), which means that the resulting solutions are almost optimal. Our algorithms ’ effectiveness is demonstrated by presenting experimental results on a variety of low level vision tasks, such as stereo matching, image restoration, image completion and optical flow estimation, as well as on synthetic problems.
A new framework for approximate labeling via graph cuts
 In International Conference on Computer Vision
, 2005
"... A new framework is presented that uses tools from duality theory of linear programming to derive graphcut based combinatorial algorithms for approximating NPhard classification problems. The derived algorithms include αexpansion graph cut techniques merely as a special case, have guaranteed optim ..."
Abstract

Cited by 35 (7 self)
 Add to MetaCart
(Show Context)
A new framework is presented that uses tools from duality theory of linear programming to derive graphcut based combinatorial algorithms for approximating NPhard classification problems. The derived algorithms include αexpansion graph cut techniques merely as a special case, have guaranteed optimality properties even in cases where αexpansion techniques fail to do so and can provide very tight perinstance suboptimality bounds in all occasions. 1
APPEARED IN THE PROCEEDINGS OF IEEE ICCV 2005 A New Framework for Approximate Labeling via Graph Cuts
"... A new framework is presented that uses tools from duality theory of linear programming to derive graphcut based combinatorial algorithms for approximating NPhard classification problems. The derived algorithms include αexpansion graph cut techniques merely as a special case, have guaranteed optim ..."
Abstract
 Add to MetaCart
(Show Context)
A new framework is presented that uses tools from duality theory of linear programming to derive graphcut based combinatorial algorithms for approximating NPhard classification problems. The derived algorithms include αexpansion graph cut techniques merely as a special case, have guaranteed optimality properties even in cases where αexpansion techniques fail to do so and can provide very tight perinstance suboptimality bounds in all occasions. 1
(u,v)∈E
, 2007
"... A note on the primaldual method for the semimetric labeling problem Recently, Komodakis et al. [6] developed the FastPD algorithm for the semimetric labeling problem, which extends the expansion move algorithm of Boykov et al. [2]. We present a slightly different derivation of the FastPD method. ..."
Abstract
 Add to MetaCart
(Show Context)
A note on the primaldual method for the semimetric labeling problem Recently, Komodakis et al. [6] developed the FastPD algorithm for the semimetric labeling problem, which extends the expansion move algorithm of Boykov et al. [2]. We present a slightly different derivation of the FastPD method. 1. Preliminaries Consider the following energy function: E(x  ¯ θ) = � ¯θu(xv) + � u∈V