Deep learning for biological image classification

Página do item simplificado
dc.contributor.authorAffonso, Carlos [UNESP]
dc.contributor.authorRossi, André Luis Debiaso [UNESP]
dc.contributor.authorVieira, Fábio Henrique Antunes [UNESP]
dc.contributor.authorde Carvalho, André Carlos Ponce de Leon Ferreira
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.contributor.institutionUniversidade de São Paulo (USP)
dc.date.accessioned2018-12-11T16:47:26Z
dc.date.available2018-12-11T16:47:26Z
dc.date.issued2017-11-01
dc.description.abstractA number of industries use human inspection to visually classify the quality of their products and the raw materials used in the production process, this process could be done automatically through digital image processing. The industries are not always interested in the most accurate technique for a given problem, but most appropriate for the expected results, there must be a balance between accuracy and computational cost. This paper investigates the classification of the quality of wood boards based on their images. For such, it compares the use of deep learning, particularly Convolutional Neural Networks, with the combination of texture-based feature extraction techniques and traditional techniques: Decision tree induction algorithms, Neural Networks, Nearest neighbors and Support vector machines. Reported studies show that Deep Learning techniques applied to image processing tasks have achieved predictive performance superior to traditional classification techniques, mainly in high complex scenarios. One of the reasons pointed out is their embedded feature extraction mechanism. Deep Learning techniques directly identify and extract features, considered by them to be relevant, in a given image dataset. However, empirical results for the image data set have shown that the texture descriptor method proposed, regardless of the strategy employed is very competitive when compared with Convolutional Neural Network for all the performed experiments. The best performance of the texture descriptor method could be caused by the nature of the image dataset. Finally are pointed out some perspectives of futures developments with the application of Active learning and Semi supervised methods.en
dc.description.affiliationUNESP - Universidade Estadual Paulista, Julio de Mesquita Filho
dc.description.affiliationICMC - USP Instituto de Ciências Matemáticas e de Computação Universidade de São Paulo, São Carlos
dc.description.affiliationUnespUNESP - Universidade Estadual Paulista, Julio de Mesquita Filho
dc.format.extent114-122
dc.identifierhttp://dx.doi.org/10.1016/j.eswa.2017.05.039
dc.identifier.citationExpert Systems with Applications, v. 85, p. 114-122.
dc.identifier.doi10.1016/j.eswa.2017.05.039
dc.identifier.file2-s2.0-85019578783.pdf
dc.identifier.issn0957-4174
dc.identifier.scopus2-s2.0-85019578783
dc.identifier.urihttp://hdl.handle.net/11449/169747
dc.language.isoeng
dc.relation.ispartofExpert Systems with Applications
dc.relation.ispartofsjr1,271
dc.rights.accessRightsAcesso aberto
dc.sourceScopus
dc.subjectDeep learning
dc.subjectImage classification
dc.subjectMachine learning
dc.subjectWood classification
dc.titleDeep learning for biological image classificationen
dc.typeArtigo
unesp.author.lattes5604829226181486[2]
unesp.author.lattes0849551883568657[1]
unesp.author.orcid0000-0001-6388-7479[2]

Arquivos

Pacote Original
Agora exibindo 1 - 1 de 1
Imagem de Miniatura
Nome:
2-s2.0-85019578783.pdf
Tamanho:
810.85 KB
Formato:
Adobe Portable Document Format
Descrição:
Baixar

Coleções

Artigos