Matches in SemOpenAlex for { <https://semopenalex.org/work/W4304080400> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W4304080400 endingPage "485" @default.
- W4304080400 startingPage "477" @default.
- W4304080400 abstract "Most of the previous network optimization works applied attention mechanism to feature maps, but neglected to embed attention into convolution kernel of the end-to-end network that is convenient to deploy. To address this issue, we present a novel attention convolution method named Kernel Attention Convolution (KAConv) to enhance the flexibility of convolution. The proposed KAConv generates different attention weights for different spatial positions of convolution kernels based on the input features, so as to dynamically adjust the parameters of convolution kernels during the forward propagation to enhance the flexibility of convolution. We decompose the convolution kernels into subkernels spatially, and generate the corresponding feature maps through which attention weights are obtained. The final refined feature maps are aggregated by the attention weighted feature maps corresponding to each subkernel. KAConv is a computationally lightweight convolution method, which not only incorporates attention into kernels but also enhances informative representations. By replacing the standard convolution with the proposed KAConv in convolutional neural networks (CNNs), the networks yield significant performance improvement. Extensive experiments on the ImageNet-1K benchmark demonstrate that KAConv outperforms existing attention mechanism-based methods. We also carry out experiments on the MS COCO and PASCAL VOC datasets to show the generalization ability of our method." @default.
- W4304080400 created "2022-10-10" @default.
- W4304080400 creator A5017562594 @default.
- W4304080400 creator A5021291356 @default.
- W4304080400 creator A5040124761 @default.
- W4304080400 creator A5052543245 @default.
- W4304080400 creator A5090454827 @default.
- W4304080400 date "2022-12-01" @default.
- W4304080400 modified "2023-09-26" @default.
- W4304080400 title "KAConv: Kernel attention convolutions" @default.
- W4304080400 cites W2101926813 @default.
- W4304080400 cites W2963045198 @default.
- W4304080400 cites W2981413347 @default.
- W4304080400 cites W2998460251 @default.
- W4304080400 cites W3105979354 @default.
- W4304080400 cites W3115819524 @default.
- W4304080400 cites W3144914540 @default.
- W4304080400 doi "https://doi.org/10.1016/j.neucom.2022.10.017" @default.
- W4304080400 hasPublicationYear "2022" @default.
- W4304080400 type Work @default.
- W4304080400 citedByCount "0" @default.
- W4304080400 crossrefType "journal-article" @default.
- W4304080400 hasAuthorship W4304080400A5017562594 @default.
- W4304080400 hasAuthorship W4304080400A5021291356 @default.
- W4304080400 hasAuthorship W4304080400A5040124761 @default.
- W4304080400 hasAuthorship W4304080400A5052543245 @default.
- W4304080400 hasAuthorship W4304080400A5090454827 @default.
- W4304080400 hasConcept C11413529 @default.
- W4304080400 hasConcept C118615104 @default.
- W4304080400 hasConcept C13280743 @default.
- W4304080400 hasConcept C134306372 @default.
- W4304080400 hasConcept C138885662 @default.
- W4304080400 hasConcept C153180895 @default.
- W4304080400 hasConcept C154945302 @default.
- W4304080400 hasConcept C177148314 @default.
- W4304080400 hasConcept C185798385 @default.
- W4304080400 hasConcept C199360897 @default.
- W4304080400 hasConcept C205649164 @default.
- W4304080400 hasConcept C2776401178 @default.
- W4304080400 hasConcept C33923547 @default.
- W4304080400 hasConcept C41008148 @default.
- W4304080400 hasConcept C41895202 @default.
- W4304080400 hasConcept C45347329 @default.
- W4304080400 hasConcept C50644808 @default.
- W4304080400 hasConcept C74193536 @default.
- W4304080400 hasConcept C75608658 @default.
- W4304080400 hasConcept C81363708 @default.
- W4304080400 hasConceptScore W4304080400C11413529 @default.
- W4304080400 hasConceptScore W4304080400C118615104 @default.
- W4304080400 hasConceptScore W4304080400C13280743 @default.
- W4304080400 hasConceptScore W4304080400C134306372 @default.
- W4304080400 hasConceptScore W4304080400C138885662 @default.
- W4304080400 hasConceptScore W4304080400C153180895 @default.
- W4304080400 hasConceptScore W4304080400C154945302 @default.
- W4304080400 hasConceptScore W4304080400C177148314 @default.
- W4304080400 hasConceptScore W4304080400C185798385 @default.
- W4304080400 hasConceptScore W4304080400C199360897 @default.
- W4304080400 hasConceptScore W4304080400C205649164 @default.
- W4304080400 hasConceptScore W4304080400C2776401178 @default.
- W4304080400 hasConceptScore W4304080400C33923547 @default.
- W4304080400 hasConceptScore W4304080400C41008148 @default.
- W4304080400 hasConceptScore W4304080400C41895202 @default.
- W4304080400 hasConceptScore W4304080400C45347329 @default.
- W4304080400 hasConceptScore W4304080400C50644808 @default.
- W4304080400 hasConceptScore W4304080400C74193536 @default.
- W4304080400 hasConceptScore W4304080400C75608658 @default.
- W4304080400 hasConceptScore W4304080400C81363708 @default.
- W4304080400 hasLocation W43040804001 @default.
- W4304080400 hasOpenAccess W4304080400 @default.
- W4304080400 hasPrimaryLocation W43040804001 @default.
- W4304080400 hasRelatedWork W2295021132 @default.
- W4304080400 hasRelatedWork W2424871898 @default.
- W4304080400 hasRelatedWork W2760085659 @default.
- W4304080400 hasRelatedWork W2929240682 @default.
- W4304080400 hasRelatedWork W2949189996 @default.
- W4304080400 hasRelatedWork W2963556241 @default.
- W4304080400 hasRelatedWork W3093612317 @default.
- W4304080400 hasRelatedWork W3106036237 @default.
- W4304080400 hasRelatedWork W3159557112 @default.
- W4304080400 hasRelatedWork W4293794540 @default.
- W4304080400 hasVolume "514" @default.
- W4304080400 isParatext "false" @default.
- W4304080400 isRetracted "false" @default.
- W4304080400 workType "article" @default.