Two-Stage Feature Generator for Handwritten Digit Classification

dc.authorid Tora, Hakan/0000-0002-0427-483X
dc.authorid Butun, Ismail/0000-0002-1723-5741
dc.authorid oztoprak, kasim/0000-0003-2483-8070
dc.authorscopusid 58671692200
dc.authorscopusid 6506642154
dc.authorscopusid 21743623400
dc.authorscopusid 35316899700
dc.authorwosid Butun, Ismail/K-1246-2015
dc.authorwosid oztoprak, kasim/U-1631-2018
dc.contributor.author Pirim, M. Altinay Gunler
dc.contributor.author Tora, Hakan
dc.contributor.author Oztoprak, Kasim
dc.contributor.author Butun, Ismail
dc.contributor.other Airframe and Powerplant Maintenance
dc.date.accessioned 2024-07-05T15:21:46Z
dc.date.available 2024-07-05T15:21:46Z
dc.date.issued 2023
dc.department Atılım University en_US
dc.department-temp [Pirim, M. Altinay Gunler] Vakifbank, TR-06200 Ankara, Turkiye; [Tora, Hakan] Atilim Univ, Dept Avion, TR-06830 Ankara, Turkiye; [Oztoprak, Kasim] Konya Food & Agr Univ, Dept Comp Engn, TR-42080 Konya, Turkiye; [Butun, Ismail] KTH Royal Inst Technol, Dept Comp Engn, SE-11428 Stockholm, Sweden; [Butun, Ismail] OSTIM Tech Univ, Dept Comp Engn, TR-06370 Ankara, Turkiye en_US
dc.description Tora, Hakan/0000-0002-0427-483X; Butun, Ismail/0000-0002-1723-5741; oztoprak, kasim/0000-0003-2483-8070 en_US
dc.description.abstract In this paper, a novel feature generator framework is proposed for handwritten digit classification. The proposed framework includes a two-stage cascaded feature generator. The first stage is based on principal component analysis (PCA), which generates projected data on principal components as features. The second one is constructed by a partially trained neural network (PTNN), which uses projected data as inputs and generates hidden layer outputs as features. The features obtained from the PCA and PTNN-based feature generator are tested on the MNIST and USPS datasets designed for handwritten digit sets. Minimum distance classifier (MDC) and support vector machine (SVM) methods are exploited as classifiers for the obtained features in association with this framework. The performance evaluation results show that the proposed framework outperforms the state-of-the-art techniques and achieves accuracies of 99.9815% and 99.9863% on the MNIST and USPS datasets, respectively. The results also show that the proposed framework achieves almost perfect accuracies, even with significantly small training data sizes. en_US
dc.description.sponsorship KTH Royal Institute of Technology en_US
dc.description.sponsorship This research was funded by KTH Royal Institute of Technology (via KTH Library's Open Access Policy). en_US
dc.identifier.citationcount 0
dc.identifier.doi 10.3390/s23208477
dc.identifier.issn 1424-8220
dc.identifier.issue 20 en_US
dc.identifier.pmid 37896570
dc.identifier.scopus 2-s2.0-85175278954
dc.identifier.uri https://doi.org/10.3390/s23208477
dc.identifier.uri https://hdl.handle.net/20.500.14411/2128
dc.identifier.volume 23 en_US
dc.identifier.wos WOS:001089765800001
dc.identifier.wosquality Q2
dc.institutionauthor Tora, Hakan
dc.language.iso en en_US
dc.publisher Mdpi en_US
dc.relation.publicationcategory Makale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanı en_US
dc.rights info:eu-repo/semantics/openAccess en_US
dc.scopus.citedbyCount 1
dc.subject minimum distance classifier en_US
dc.subject neural network en_US
dc.subject principal component analysis en_US
dc.subject support vector machine en_US
dc.subject pattern recognition en_US
dc.subject soft sensor en_US
dc.title Two-Stage Feature Generator for Handwritten Digit Classification en_US
dc.type Article en_US
dc.wos.citedbyCount 0
dspace.entity.type Publication
relation.isAuthorOfPublication 3b369df4-6f40-4e7f-9021-94de8b562a0d
relation.isAuthorOfPublication.latestForDiscovery 3b369df4-6f40-4e7f-9021-94de8b562a0d
relation.isOrgUnitOfPublication 0ad0b148-c2aa-44e7-8f0a-53ab5c8406d5
relation.isOrgUnitOfPublication.latestForDiscovery 0ad0b148-c2aa-44e7-8f0a-53ab5c8406d5

Files

Collections