Publication:
The effect of class-weighted penalization in deep neural networks for multi-class cell segmentation

dc.contributor.coauthorAydin, Musa
dc.contributor.coauthorKus, Zeki
dc.contributor.coauthorKiraz, Berna
dc.contributor.coauthorHosavci, Reyhan
dc.contributor.departmentDepartment of Physics
dc.contributor.kuauthorKiraz, Alper
dc.contributor.schoolcollegeinstituteCollege of Sciences
dc.date.accessioned2025-03-06T20:57:15Z
dc.date.issued2024
dc.description.abstractDeep learning networks give successful results in many areas, but their complexity leads to problems such as overfitting. Many approaches have been proposed to solve this problem, and class-based penalization has been one of the methods that have yielded successful results. With class-based penalization, it has become possible to increase the prediction performance and improve the model's generalization capability, especially in cases with class imbalance. This study investigates the effect of class-based penalization on the multiclass cell segmentation problem. Two deep neural network models (Resnet18, EfficientNet) are tested with 6 different configurations created for class-based penalization, and the results are compared. The experimental studies show the relationship between class-based loss penalties and multiclass segmentation/classification performance. The results show that class-based penalization improves the total performance of EfficientNet and Resnet18 networks by 11.82(C2+C3) and 12.79(C2+C3) points, respectively. It is shown that the proposed method can improve the prediction performance without increasing the model complexity.
dc.description.indexedbyWOS
dc.description.indexedbyScopus
dc.description.publisherscopeInternational
dc.description.sponsoredbyTubitakEuN/A
dc.identifier.doi10.1109/SIU61531.2024.10601040
dc.identifier.isbn9798350388978
dc.identifier.isbn9798350388961
dc.identifier.issn2165-0608
dc.identifier.quartileN/A
dc.identifier.scopus2-s2.0-85200860403
dc.identifier.urihttps://doi.org/10.1109/SIU61531.2024.10601040
dc.identifier.urihttps://hdl.handle.net/20.500.14288/27170
dc.identifier.wos1297894700253
dc.keywordsCell segmentation
dc.keywordsClassification
dc.keywordsClass weighted penalization
dc.language.isotur
dc.publisherIEEE
dc.relation.ispartof32ND IEEE SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE, SIU 2024
dc.subjectComputer science
dc.subjectElectrical and electronic
dc.subjectTelecommunications
dc.titleThe effect of class-weighted penalization in deep neural networks for multi-class cell segmentation
dc.title.alternativeÇok sınıflı hücre bölütleme için derin sinir ağlarında sınıf ağırlıklı cezalandırmanın etkisi
dc.typeConference Proceeding
dspace.entity.typePublication
local.contributor.kuauthorKiraz, Alper
local.publication.orgunit1College of Sciences
local.publication.orgunit2Department of Physics
relation.isOrgUnitOfPublicationc43d21f0-ae67-4f18-a338-bcaedd4b72a4
relation.isOrgUnitOfPublication.latestForDiscoveryc43d21f0-ae67-4f18-a338-bcaedd4b72a4
relation.isParentOrgUnitOfPublicationaf0395b0-7219-4165-a909-7016fa30932d
relation.isParentOrgUnitOfPublication.latestForDiscoveryaf0395b0-7219-4165-a909-7016fa30932d

Files