Faster α-expansion via dynamic programming and image partitioning

dc.contributor.authorFontinele, Jefferson
dc.contributor.authorMendonca, Marcelo
dc.contributor.authorRuiz, Marco
dc.contributor.authorPapa, Joao [UNESP]
dc.contributor.authorOliveira, Luciano
dc.contributor.institutionUniversidade Federal da Bahia (UFBA)
dc.contributor.institutionVORTEX-CoLab
dc.contributor.institutionUniversidade Estadual Paulista (UNESP)
dc.date.accessioned2022-04-28T19:29:29Z
dc.date.available2022-04-28T19:29:29Z
dc.date.issued2020-07-01
dc.description.abstractImage segmentation is the task of assigning a label to each image pixel. When the number of labels is greater than two (multi-label) the segmentation can be modelled as a multi-cut problem in graphs. In the general case, finding the minimum cut in a graph is an NP-hard problem, in which improving the results concerning time and quality is a major challenge. This paper addresses the multi-label problem applied in interactive image segmentation. The proposed approach makes use of dynamic programming to initialize an α-expansion, thus reducing its runtime, while keeping the Dice-score measure in an interactive segmentation task. Over BSDS data set, the proposed algorithm was approximately 51.2% faster than its standard counterpart, 36.2% faster than Fast Primal-Dual (FastPD) and 10.5 times faster than quadratic pseudo-boolean optimization (QBPO) optimizers, while preserving the same segmentation quality.en
dc.description.affiliationFederal University of Bahia Intelligent Vision Research Lab
dc.description.affiliationVORTEX-CoLab
dc.description.affiliationSão Paulo State University
dc.description.affiliationUnespSão Paulo State University
dc.identifierhttp://dx.doi.org/10.1109/IJCNN48605.2020.9207032
dc.identifier.citationProceedings of the International Joint Conference on Neural Networks.
dc.identifier.doi10.1109/IJCNN48605.2020.9207032
dc.identifier.scopus2-s2.0-85093828760
dc.identifier.urihttp://hdl.handle.net/11449/221592
dc.language.isoeng
dc.relation.ispartofProceedings of the International Joint Conference on Neural Networks
dc.sourceScopus
dc.subjectdynamic programming
dc.subjectimage segmentation
dc.subjectmulti-label
dc.subjectα-expansion
dc.titleFaster α-expansion via dynamic programming and image partitioningen
dc.typeTrabalho apresentado em evento

Arquivos

Coleções