Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287725747> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4287725747 abstract "Capsule Networks (CapsNets) have demonstrated to be a promising alternative to Convolutional Neural Networks (CNNs). However, they often fall short of state-of-the-art accuracies on large-scale high-dimensional datasets. We propose a Detail-Oriented Capsule Network (DECAPS) that combines the strength of CapsNets with several novel techniques to boost its classification accuracies. First, DECAPS uses an Inverted Dynamic Routing (IDR) mechanism to group lower-level capsules into heads before sending them to higher-level capsules. This strategy enables capsules to selectively attend to small but informative details within the data which may be lost during pooling operations in CNNs. Second, DECAPS employs a Peekaboo training procedure, which encourages the network to focus on fine-grained information through a second-level attention scheme. Finally, the distillation process improves the robustness of DECAPS by averaging over the original and attended image region predictions. We provide extensive experiments on the CheXpert and RSNA Pneumonia datasets to validate the effectiveness of DECAPS. Our networks achieve state-of-the-art accuracies not only in classification (increasing the average area under ROC curves from 87.24% to 92.82% on the CheXpert dataset) but also in the weakly-supervised localization of diseased areas (increasing average precision from 41.7% to 80% for the RSNA Pneumonia detection dataset)." @default.
- W4287725747 created "2022-07-26" @default.
- W4287725747 creator A5060060936 @default.
- W4287725747 creator A5077363470 @default.
- W4287725747 creator A5080296240 @default.
- W4287725747 creator A5085740219 @default.
- W4287725747 date "2020-07-08" @default.
- W4287725747 modified "2023-09-29" @default.
- W4287725747 title "DECAPS: Detail-Oriented Capsule Networks" @default.
- W4287725747 doi "https://doi.org/10.48550/arxiv.2007.05343" @default.
- W4287725747 hasPublicationYear "2020" @default.
- W4287725747 type Work @default.
- W4287725747 citedByCount "0" @default.
- W4287725747 crossrefType "posted-content" @default.
- W4287725747 hasAuthorship W4287725747A5060060936 @default.
- W4287725747 hasAuthorship W4287725747A5077363470 @default.
- W4287725747 hasAuthorship W4287725747A5080296240 @default.
- W4287725747 hasAuthorship W4287725747A5085740219 @default.
- W4287725747 hasBestOaLocation W42877257471 @default.
- W4287725747 hasConcept C104317684 @default.
- W4287725747 hasConcept C104954878 @default.
- W4287725747 hasConcept C119857082 @default.
- W4287725747 hasConcept C124101348 @default.
- W4287725747 hasConcept C153180895 @default.
- W4287725747 hasConcept C154945302 @default.
- W4287725747 hasConcept C185592680 @default.
- W4287725747 hasConcept C24856439 @default.
- W4287725747 hasConcept C31258907 @default.
- W4287725747 hasConcept C41008148 @default.
- W4287725747 hasConcept C55493867 @default.
- W4287725747 hasConcept C63479239 @default.
- W4287725747 hasConcept C70437156 @default.
- W4287725747 hasConcept C74172769 @default.
- W4287725747 hasConcept C81363708 @default.
- W4287725747 hasConcept C9659607 @default.
- W4287725747 hasConceptScore W4287725747C104317684 @default.
- W4287725747 hasConceptScore W4287725747C104954878 @default.
- W4287725747 hasConceptScore W4287725747C119857082 @default.
- W4287725747 hasConceptScore W4287725747C124101348 @default.
- W4287725747 hasConceptScore W4287725747C153180895 @default.
- W4287725747 hasConceptScore W4287725747C154945302 @default.
- W4287725747 hasConceptScore W4287725747C185592680 @default.
- W4287725747 hasConceptScore W4287725747C24856439 @default.
- W4287725747 hasConceptScore W4287725747C31258907 @default.
- W4287725747 hasConceptScore W4287725747C41008148 @default.
- W4287725747 hasConceptScore W4287725747C55493867 @default.
- W4287725747 hasConceptScore W4287725747C63479239 @default.
- W4287725747 hasConceptScore W4287725747C70437156 @default.
- W4287725747 hasConceptScore W4287725747C74172769 @default.
- W4287725747 hasConceptScore W4287725747C81363708 @default.
- W4287725747 hasConceptScore W4287725747C9659607 @default.
- W4287725747 hasLocation W42877257471 @default.
- W4287725747 hasOpenAccess W4287725747 @default.
- W4287725747 hasPrimaryLocation W42877257471 @default.
- W4287725747 hasRelatedWork W2424871898 @default.
- W4287725747 hasRelatedWork W2514274290 @default.
- W4287725747 hasRelatedWork W2517027266 @default.
- W4287725747 hasRelatedWork W2758063741 @default.
- W4287725747 hasRelatedWork W2788663687 @default.
- W4287725747 hasRelatedWork W2792080776 @default.
- W4287725747 hasRelatedWork W2921836287 @default.
- W4287725747 hasRelatedWork W2943085063 @default.
- W4287725747 hasRelatedWork W2951391129 @default.
- W4287725747 hasRelatedWork W2969680539 @default.
- W4287725747 isParatext "false" @default.
- W4287725747 isRetracted "false" @default.
- W4287725747 workType "article" @default.