LP-relaxation of binarized energy minimization

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F08%3A03150825" target="_blank" >RIV/68407700:21230/08:03150825 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

LP-relaxation of binarized energy minimization

Popis výsledku v původním jazyce

We address the problem of energy minimization, which is (1) generally NP-complete and (2) involves many discrete variables - commonly a 2D array of them, arising from an MRF model. One of the approaches to the problem is to formulate it as integer linearprogramming and relax integrality constraints. However this can be done in a number of possible ways. One, widely applied previously (LP-1) [19, 13, 4, 22, 9, 23], appears to lead to a large-scale linear program which is not practical to solve with general LP methods. A number of algorithms were developed which attempt to solve the problem exploiting its structure [14, 23, 22, 9], however their common drawback is that they may converge to a suboptimal point. The other LP relaxation we consider here isconstructed by (1) refor- mulating the optimization problem in the form of a function of binary vari- ables [18], and (2) applying the roof duality relaxation [6] to the reformulated problem. We refer to the resulting relaxation as LP-2.

Název v anglickém jazyce

LP-relaxation of binarized energy minimization

Popis výsledku anglicky

We address the problem of energy minimization, which is (1) generally NP-complete and (2) involves many discrete variables - commonly a 2D array of them, arising from an MRF model. One of the approaches to the problem is to formulate it as integer linearprogramming and relax integrality constraints. However this can be done in a number of possible ways. One, widely applied previously (LP-1) [19, 13, 4, 22, 9, 23], appears to lead to a large-scale linear program which is not practical to solve with general LP methods. A number of algorithms were developed which attempt to solve the problem exploiting its structure [14, 23, 22, 9], however their common drawback is that they may converge to a suboptimal point. The other LP relaxation we consider here isconstructed by (1) refor- mulating the optimization problem in the form of a function of binary vari- ables [18], and (2) applying the roof duality relaxation [6] to the reformulated problem. We refer to the resulting relaxation as LP-2.

Druh

O - Ostatní výsledky

CEP obor

JD - Využití počítačů, robotika a její aplikace

OECD FORD obor

—

Projekt

<a href="/cs/project/7E08031" target="_blank" >7E08031: Dynamic Interactive Perception-action Learning in Cognitive Systems</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>Z - Vyzkumny zamer (s odkazem do CEZ)

Rok uplatnění

2008

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů