Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912442366> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W2912442366 endingPage "787" @default.
- W2912442366 startingPage "781" @default.
- W2912442366 abstract "摘要 针对基于卷积神经网络的图像识别采用随机初始化网络权值的方法易收敛到局部最优值的问题, 该文提出了一种结合无监督和有监督学习的网络权值预训练算法。融合零成分分析白化与深度信念网络预学习得到的特征, 对卷积神经网络权值进行初始化; 通过卷积、池化等操作, 对训练样本进行特征提取并使用全连接网络对特征进行分类; 计算分类损失函数并优化网络参数。在公开图像数据库中进行了大量实验, 与公开最佳算法比较, 该算法在MNIST中的识别错误率降低了0.1%, 在Caltech101中的分类准确率提升了0.56%, 验证了该算法优于现有算法。" @default.
- W2912442366 created "2019-02-21" @default.
- W2912442366 creator A5012301761 @default.
- W2912442366 creator A5047904684 @default.
- W2912442366 creator A5055170145 @default.
- W2912442366 creator A5084899203 @default.
- W2912442366 creator A5087041063 @default.
- W2912442366 date "2018-09-19" @default.
- W2912442366 modified "2023-09-24" @default.
- W2912442366 title "Image recognition and classification by deep belief-convolutional neural networks" @default.
- W2912442366 doi "https://doi.org/10.16511/j.cnki.qhdxxb.2018.22.034" @default.
- W2912442366 hasPublicationYear "2018" @default.
- W2912442366 type Work @default.
- W2912442366 sameAs 2912442366 @default.
- W2912442366 citedByCount "0" @default.
- W2912442366 crossrefType "journal-article" @default.
- W2912442366 hasAuthorship W2912442366A5012301761 @default.
- W2912442366 hasAuthorship W2912442366A5047904684 @default.
- W2912442366 hasAuthorship W2912442366A5055170145 @default.
- W2912442366 hasAuthorship W2912442366A5084899203 @default.
- W2912442366 hasAuthorship W2912442366A5087041063 @default.
- W2912442366 hasConcept C108583219 @default.
- W2912442366 hasConcept C115961682 @default.
- W2912442366 hasConcept C153180895 @default.
- W2912442366 hasConcept C154945302 @default.
- W2912442366 hasConcept C31972630 @default.
- W2912442366 hasConcept C41008148 @default.
- W2912442366 hasConcept C50644808 @default.
- W2912442366 hasConcept C75294576 @default.
- W2912442366 hasConcept C81363708 @default.
- W2912442366 hasConceptScore W2912442366C108583219 @default.
- W2912442366 hasConceptScore W2912442366C115961682 @default.
- W2912442366 hasConceptScore W2912442366C153180895 @default.
- W2912442366 hasConceptScore W2912442366C154945302 @default.
- W2912442366 hasConceptScore W2912442366C31972630 @default.
- W2912442366 hasConceptScore W2912442366C41008148 @default.
- W2912442366 hasConceptScore W2912442366C50644808 @default.
- W2912442366 hasConceptScore W2912442366C75294576 @default.
- W2912442366 hasConceptScore W2912442366C81363708 @default.
- W2912442366 hasIssue "9" @default.
- W2912442366 hasLocation W29124423661 @default.
- W2912442366 hasOpenAccess W2912442366 @default.
- W2912442366 hasPrimaryLocation W29124423661 @default.
- W2912442366 hasRelatedWork W2406228418 @default.
- W2912442366 hasRelatedWork W2612125541 @default.
- W2912442366 hasRelatedWork W2626673250 @default.
- W2912442366 hasRelatedWork W2734592973 @default.
- W2912442366 hasRelatedWork W2742520671 @default.
- W2912442366 hasRelatedWork W2923183175 @default.
- W2912442366 hasRelatedWork W2945390381 @default.
- W2912442366 hasRelatedWork W2955313527 @default.
- W2912442366 hasRelatedWork W2955723949 @default.
- W2912442366 hasRelatedWork W2971745841 @default.
- W2912442366 hasRelatedWork W2988831283 @default.
- W2912442366 hasRelatedWork W3016247907 @default.
- W2912442366 hasRelatedWork W3042976221 @default.
- W2912442366 hasRelatedWork W3082182714 @default.
- W2912442366 hasRelatedWork W3087406541 @default.
- W2912442366 hasRelatedWork W3127265066 @default.
- W2912442366 hasRelatedWork W3195081136 @default.
- W2912442366 hasRelatedWork W3200934184 @default.
- W2912442366 hasRelatedWork W3206325217 @default.
- W2912442366 hasRelatedWork W3179806503 @default.
- W2912442366 hasVolume "58" @default.
- W2912442366 isParatext "false" @default.
- W2912442366 isRetracted "false" @default.
- W2912442366 magId "2912442366" @default.
- W2912442366 workType "article" @default.