Deep Graph Matching via Blackbox Differentiation of
Combinatorial Solvers

D 2020

Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers

ROLÍNEK, Michal; Paul SWOBODA; Dominik ZIETLOW; Anselm PAULUS; Vít MUSIL et al.

Základní údaje

Originální název

Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers

Autoři

ROLÍNEK, Michal; Paul SWOBODA; Dominik ZIETLOW; Anselm PAULUS; Vít MUSIL a Georg MARTIUS

Vydání

12373 LNCS. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), od s. 407-424, 18 s. 2020

Nakladatel

Springer Science and Business Media Deutschland GmbH

Další údaje

Jazyk

angličtina

Typ výsledku

Stať ve sborníku

Obor

10101 Pure mathematics

Utajení

není předmětem státního či obchodního tajemství

Forma vydání

elektronická verze "online"

Odkazy

Springer

Označené pro přenos do RIV

Organizační jednotka

Fakulta informatiky

ISBN

978-3-030-58603-4

DOI

https://doi.org/10.1007/978-3-030-58604-1_25

EID Scopus

2-s2.0-85097044661

Klíčová slova anglicky

Combinatorial optimization; Deep graph matching; Keypoint correspondence

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 7. 11. 2024 12:00, RNDr. Vít Musil, Ph.D.

Anotace

V originále

Building on recent progress at the intersection of combinatorial optimization and deep learning, we propose an end-to-end trainable architecture for deep graph matching that contains unmodified combinatorial solvers. Using the presence of heavily optimized combinatorial solvers together with some improvements in architecture design, we advance state-of-the-art on deep graph matching benchmarks for keypoint correspondence. In addition, we highlight the conceptual advantages of incorporating solvers into deep learning architectures, such as the possibility of post-processing with a strong multi-graph matching solver or the indifference to changes in the training setting. Finally, we propose two new challenging experimental setups.

Citovat

ROLÍNEK, Michal; Paul SWOBODA; Dominik ZIETLOW; Anselm PAULUS; Vít MUSIL a Georg MARTIUS. Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers. Online. In Vedaldi A.Bischof H.Brox T.Frahm J. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 12373 LNCS. Springer Science and Business Media Deutschland GmbH, 2020, s. 407-424. ISBN 978-3-030-58603-4. Dostupné z: https://doi.org/10.1007/978-3-030-58604-1_25.

@inproceedings{2449839,
   author = {Rolínek, Michal and Swoboda, Paul and Zietlow, Dominik and Paulus, Anselm and Musil, Vít and Martius, Georg},
   booktitle = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
   doi = {https://doi.org/10.1007/978-3-030-58604-1_25},
   edition = {12373 LNCS},
   editor = {Vedaldi A.Bischof H.Brox T.Frahm J.},
   keywords = {Combinatorial optimization; Deep graph matching; Keypoint correspondence},
   howpublished = {elektronická verze "online"},
   language = {eng},
   isbn = {978-3-030-58603-4},
   pages = {407-424},
   publisher = {Springer Science and Business Media Deutschland GmbH},
   title = {Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers},
   url = {https://link.springer.com/chapter/10.1007/978-3-030-58604-1_25},
   year = {2020}
}

TY  - CONF
ID  - 2449839
AU  - Rolínek, Michal - Swoboda, Paul - Zietlow, Dominik - Paulus, Anselm - Musil, Vít - Martius, Georg
PY  - 2020
TI  - Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers
PB  - Springer Science and Business Media Deutschland GmbH
SN  - 9783030586034
KW  - Combinatorial optimization
KW  - Deep graph matching
KW  - Keypoint correspondence
UR  - https://link.springer.com/chapter/10.1007/978-3-030-58604-1_25
N2  - Building on recent progress at the intersection of combinatorial optimization and deep learning, we propose an end-to-end trainable architecture for deep graph matching that contains unmodified combinatorial solvers. Using the presence of heavily optimized combinatorial solvers together with some improvements in architecture design, we advance state-of-the-art on deep graph matching benchmarks for keypoint correspondence. In addition, we highlight the conceptual advantages of incorporating solvers into deep learning architectures, such as the possibility of post-processing with a strong multi-graph matching solver or the indifference to changes in the training setting. Finally, we propose two new challenging experimental setups.
ER  -

ROLÍNEK, Michal; Paul SWOBODA; Dominik ZIETLOW; Anselm PAULUS; Vít MUSIL a Georg MARTIUS. Deep Graph Matching via Blackbox Differentiation of Combinatorial Solvers. Online. In Vedaldi A.Bischof H.Brox T.Frahm J. \textit{Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}. 12373 LNCS. Springer Science and Business Media Deutschland GmbH, 2020, s.~407-424. ISBN~978-3-030-58603-4. Dostupné z: https://doi.org/10.1007/978-3-030-58604-1\_{}25.

Přehled o publikaci