Feature-Based Multi-Object Tracking With Maximally One Object per Class

dc.contributor.authorKrejčí, Jan
dc.contributor.authorStraka, Ondřej
dc.contributor.authorVyskočil, Jiří
dc.contributor.authorJiřík, Miroslav
dc.contributor.authorDahmen, Uta
dc.date.accessioned2023-01-02T11:00:10Z
dc.date.available2023-01-02T11:00:10Z
dc.date.issued2022
dc.description.abstract-translatedThis paper deals with the problem of tracking multiple objects, in which each object is known to belong to a unique class. We follow the tracking by detection paradigm and assume that the object detector provides scores in addition to each detection. The problem is tackled as simultaneous classification and tracking using random finite sets. Inspired by the multi-Bernoulli mixture (MBM) filter, we propose a solution to the problem by modifying the target birth process. To simplify the implementation and to mitigate the computational costs, we develop tractable solutions with linear complexity. The algorithms are subsequently used for visual tracking of surgical instruments. As a by-product, we derive the prediction step of the Bernoulli filter using the probability generating functionals (PGFLs).en
dc.format8 s.cs
dc.format.mimetypeapplication/pdf
dc.identifier.citationKREJČÍ, J. STRAKA, O. VYSKOČIL, J. JIŘÍK, M. DAHMEN, U. Feature-Based Multi-Object Tracking With Maximally One Object per Class. In Proceedings of the 25th International Conference on Information Fusion, FUSION 2022. New York: IEEE, 2022. s. 1-8. ISBN: 978-1-73774-972-1 , ISSN: neuvedenocs
dc.identifier.document-number855689000104
dc.identifier.doi10.23919/FUSION49751.2022.9841332
dc.identifier.isbn978-1-73774-972-1
dc.identifier.issnneuvedeno
dc.identifier.obd43936987
dc.identifier.uri2-s2.0-85136562114
dc.identifier.urihttp://hdl.handle.net/11025/50805
dc.language.isoenen
dc.project.IDSGS-2019-027/Inteligentní metody strojového vnímání a porozumění 4cs
dc.project.IDEF19_073/0016931/Zvyšování kvality interních grantových schémat na ZČUcs
dc.project.IDSGS-2022-022/Rozvoj a využití kybernetických systémů identifikace, diagnostiky a řízení 5cs
dc.project.IDLM2018140/E-infrastruktura CZcs
dc.project.IDIDEG-2021-009/Multiple object visual trackingcs
dc.publisherIEEEen
dc.relation.ispartofseriesProceedings of the 25th International Conference on Information Fusion, FUSION 2022en
dc.rightsPlný text je přístupný v rámci univerzity přihlášeným uživatelům.cs
dc.rights© International Society of Information Fusionen
dc.rights.accessrestrictedAccessen
dc.subject.translatedVisual tracking, random finite sets, multi-Bernoulli mixture, joint tracking and classificationen
dc.titleFeature-Based Multi-Object Tracking With Maximally One Object per Classen
dc.typekonferenční příspěvekcs
dc.typeConferenceObjecten
dc.type.statusPeer-revieweden
dc.type.versionpublishedVersionen

Files