Classification of Pollen Images with Structural Characteristics

dc.authoridCaliskan, Abidin/0000-0001-5039-6400
dc.contributor.authorErez, M. Emre
dc.contributor.authorKaya, Yilmaz
dc.contributor.authorCaliskan, Abidin
dc.date.accessioned2024-12-24T19:23:55Z
dc.date.available2024-12-24T19:23:55Z
dc.date.issued2013
dc.departmentSiirt Üniversitesi
dc.description21st Signal Processing and Communications Applications Conference (SIU) -- APR 24-26, 2013 -- CYPRUS
dc.description.abstractIn this study, a computer vision system has been developed to separate the pollen grains of plants according to their taxonomic categories without the help of an expert person. Pollen grains have a complex three-dimensional structure however they can be distinguished from one to another with their specific features. In the research, for the classification of pollen images the local edge patterns (LEP) were used. The proposed system is consists of three stages. At first Stage, Sobel edge detection algorithm was applied to pollen images to obtained new images that have prominent structural features. At the second stage LEP features were obtained and at the last stage the classification process was performed by machine learning methods by LEP features. The 98.48% classification success were obtained by LEP features.
dc.identifier.isbn978-1-4673-5563-6
dc.identifier.isbn978-1-4673-5562-9
dc.identifier.issn2165-0608
dc.identifier.urihttps://hdl.handle.net/20.500.12604/5748
dc.identifier.wosWOS:000325005300172
dc.identifier.wosqualityN/A
dc.indekslendigikaynakWeb of Science
dc.language.isotr
dc.publisherIEEE
dc.relation.ispartof2013 21st Signal Processing and Communications Applications Conference (Siu)
dc.relation.publicationcategoryKonferans Öğesi - Uluslararası - Kurum Öğretim Elemanı
dc.rightsinfo:eu-repo/semantics/closedAccess
dc.snmzKA_20241222
dc.subjectPollen
dc.subjectPollen identification
dc.subjectlocal binary pattern
dc.subjectstructural features
dc.titleClassification of Pollen Images with Structural Characteristics
dc.typeConference Object

Dosyalar