Bayesian Network for algorithm selection: Real-world hierarchy for nodes reduction

Martin Lukac, Michitaka Kameyama

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

In order to obtain the best result in image understanding it is desirable to select the best algorithm on a case by case basis. An algorithm can be selected using only image features, however such selected algorithms will often generate errors due to occlusion, shadows and other environmental conditions. To avoid such errors, it is necessary to understand processing errors on a symbolic level. Using symbolic information to determine the best algorithm is however difficult task because the possible combinations of elements and environmental conditions are almost infinite. Consequently it is impossible to predict best algorithm for all possible combinations of objects, environment conditions and context variations. In this paper we investigate selection of algorithms using symbolic image description and the determination of algorithms' error from high level image description. The proposed method transforms and minimize the high level information contained in the symbolic image description in such manner that will preserve the algorithm selection quality. The transformation takes a high level information label and transforms it into a set of generic features while the minimization uses hierarchy to reduce the specific nature of the information. Both methods of information reduction are used in a Bayesian Network because a BN is well known for using the generalization and hierarchy. As is shown in this paper, such representation efficiently reduces the fine grain high-level symbolic description to a coarser-grain hierarchy that preserves the selection quality but reduces the number of nodes.

Original languageEnglish
Title of host publication2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013
PublisherIEEE Computer Society
Pages69-74
Number of pages6
DOIs
Publication statusPublished - 2013
Externally publishedYes
Event2013 International Joint Conference on Awareness Science and Technology, iCAST 2013 and 6th International Conference on Ubi-Media Computing, UMEDIA 2013 - Aizuwakamatsu, Japan
Duration: Nov 2 2013Nov 4 2013

Other

Other2013 International Joint Conference on Awareness Science and Technology, iCAST 2013 and 6th International Conference on Ubi-Media Computing, UMEDIA 2013
CountryJapan
CityAizuwakamatsu
Period11/2/1311/4/13

Fingerprint

Bayesian networks
Image understanding
Labels
Processing

ASJC Scopus subject areas

  • Software

Cite this

Lukac, M., & Kameyama, M. (2013). Bayesian Network for algorithm selection: Real-world hierarchy for nodes reduction. In 2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013 (pp. 69-74). [6765411] IEEE Computer Society. https://doi.org/10.1109/ICAwST.2013.6765411

Bayesian Network for algorithm selection : Real-world hierarchy for nodes reduction. / Lukac, Martin; Kameyama, Michitaka.

2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013. IEEE Computer Society, 2013. p. 69-74 6765411.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Lukac, M & Kameyama, M 2013, Bayesian Network for algorithm selection: Real-world hierarchy for nodes reduction. in 2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013., 6765411, IEEE Computer Society, pp. 69-74, 2013 International Joint Conference on Awareness Science and Technology, iCAST 2013 and 6th International Conference on Ubi-Media Computing, UMEDIA 2013, Aizuwakamatsu, Japan, 11/2/13. https://doi.org/10.1109/ICAwST.2013.6765411
Lukac M, Kameyama M. Bayesian Network for algorithm selection: Real-world hierarchy for nodes reduction. In 2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013. IEEE Computer Society. 2013. p. 69-74. 6765411 https://doi.org/10.1109/ICAwST.2013.6765411
Lukac, Martin ; Kameyama, Michitaka. / Bayesian Network for algorithm selection : Real-world hierarchy for nodes reduction. 2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013. IEEE Computer Society, 2013. pp. 69-74
@inproceedings{85ca0e92331b412bba3d46efd0d24121,
title = "Bayesian Network for algorithm selection: Real-world hierarchy for nodes reduction",
abstract = "In order to obtain the best result in image understanding it is desirable to select the best algorithm on a case by case basis. An algorithm can be selected using only image features, however such selected algorithms will often generate errors due to occlusion, shadows and other environmental conditions. To avoid such errors, it is necessary to understand processing errors on a symbolic level. Using symbolic information to determine the best algorithm is however difficult task because the possible combinations of elements and environmental conditions are almost infinite. Consequently it is impossible to predict best algorithm for all possible combinations of objects, environment conditions and context variations. In this paper we investigate selection of algorithms using symbolic image description and the determination of algorithms' error from high level image description. The proposed method transforms and minimize the high level information contained in the symbolic image description in such manner that will preserve the algorithm selection quality. The transformation takes a high level information label and transforms it into a set of generic features while the minimization uses hierarchy to reduce the specific nature of the information. Both methods of information reduction are used in a Bayesian Network because a BN is well known for using the generalization and hierarchy. As is shown in this paper, such representation efficiently reduces the fine grain high-level symbolic description to a coarser-grain hierarchy that preserves the selection quality but reduces the number of nodes.",
author = "Martin Lukac and Michitaka Kameyama",
year = "2013",
doi = "10.1109/ICAwST.2013.6765411",
language = "English",
pages = "69--74",
booktitle = "2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013",
publisher = "IEEE Computer Society",
address = "United States",

}

TY - GEN

T1 - Bayesian Network for algorithm selection

T2 - Real-world hierarchy for nodes reduction

AU - Lukac, Martin

AU - Kameyama, Michitaka

PY - 2013

Y1 - 2013

N2 - In order to obtain the best result in image understanding it is desirable to select the best algorithm on a case by case basis. An algorithm can be selected using only image features, however such selected algorithms will often generate errors due to occlusion, shadows and other environmental conditions. To avoid such errors, it is necessary to understand processing errors on a symbolic level. Using symbolic information to determine the best algorithm is however difficult task because the possible combinations of elements and environmental conditions are almost infinite. Consequently it is impossible to predict best algorithm for all possible combinations of objects, environment conditions and context variations. In this paper we investigate selection of algorithms using symbolic image description and the determination of algorithms' error from high level image description. The proposed method transforms and minimize the high level information contained in the symbolic image description in such manner that will preserve the algorithm selection quality. The transformation takes a high level information label and transforms it into a set of generic features while the minimization uses hierarchy to reduce the specific nature of the information. Both methods of information reduction are used in a Bayesian Network because a BN is well known for using the generalization and hierarchy. As is shown in this paper, such representation efficiently reduces the fine grain high-level symbolic description to a coarser-grain hierarchy that preserves the selection quality but reduces the number of nodes.

AB - In order to obtain the best result in image understanding it is desirable to select the best algorithm on a case by case basis. An algorithm can be selected using only image features, however such selected algorithms will often generate errors due to occlusion, shadows and other environmental conditions. To avoid such errors, it is necessary to understand processing errors on a symbolic level. Using symbolic information to determine the best algorithm is however difficult task because the possible combinations of elements and environmental conditions are almost infinite. Consequently it is impossible to predict best algorithm for all possible combinations of objects, environment conditions and context variations. In this paper we investigate selection of algorithms using symbolic image description and the determination of algorithms' error from high level image description. The proposed method transforms and minimize the high level information contained in the symbolic image description in such manner that will preserve the algorithm selection quality. The transformation takes a high level information label and transforms it into a set of generic features while the minimization uses hierarchy to reduce the specific nature of the information. Both methods of information reduction are used in a Bayesian Network because a BN is well known for using the generalization and hierarchy. As is shown in this paper, such representation efficiently reduces the fine grain high-level symbolic description to a coarser-grain hierarchy that preserves the selection quality but reduces the number of nodes.

UR - http://www.scopus.com/inward/record.url?scp=84899429308&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84899429308&partnerID=8YFLogxK

U2 - 10.1109/ICAwST.2013.6765411

DO - 10.1109/ICAwST.2013.6765411

M3 - Conference contribution

AN - SCOPUS:84899429308

SP - 69

EP - 74

BT - 2013 International Joint Conference on Awareness Science and Technology and Ubi-Media Computing: Can We Realize Awareness via Ubi-Media?, iCAST 2013 and UMEDIA 2013

PB - IEEE Computer Society

ER -