Matches in SemOpenAlex for { <https://semopenalex.org/work/W3126366999> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W3126366999 endingPage "322" @default.
- W3126366999 startingPage "312" @default.
- W3126366999 abstract "Contemporary segmentation methods are usually based on deep fully convolutional networks (FCNs). However, the layer-by-layer convolutions with a growing receptive field is not good at capturing long-range contexts such as lane markers in the scene. In this paper, we address this issue by designing a distillation method that exploits label structure when training segmentation network. The intuition is that the ground-truth lane annotations themselves exhibit internal structure. We broadcast the structure hints throughout a teacher network, i.e., we train a teacher network that consumes a lane label map as input and attempts to replicate it as output. Then, the attention maps of the teacher network are adopted as supervisors of the student segmentation network. The teacher network, with label structure information embedded, knows distinctly where the convolution layers should pay visual attention into. The proposed method is named as Label-guided Attention Distillation (LGAD). It turns out that the student network learns significantly better with LGAD than when learning alone. As the teacher network is deprecated after training, our method do not increase the inference time. Note that LGAD can be easily incorporated in any lane segmentation network." @default.
- W3126366999 created "2021-02-15" @default.
- W3126366999 creator A5074956841 @default.
- W3126366999 creator A5081147343 @default.
- W3126366999 date "2021-05-01" @default.
- W3126366999 modified "2023-09-25" @default.
- W3126366999 title "Label-guided Attention Distillation for lane segmentation" @default.
- W3126366999 cites W114517082 @default.
- W3126366999 cites W1980026261 @default.
- W3126366999 cites W2117539524 @default.
- W3126366999 cites W2277132981 @default.
- W3126366999 cites W2294873883 @default.
- W3126366999 cites W2395611524 @default.
- W3126366999 cites W2412782625 @default.
- W3126366999 cites W2780740184 @default.
- W3126366999 cites W2799406003 @default.
- W3126366999 cites W2963856865 @default.
- W3126366999 cites W2963881378 @default.
- W3126366999 doi "https://doi.org/10.1016/j.neucom.2021.01.100" @default.
- W3126366999 hasPublicationYear "2021" @default.
- W3126366999 type Work @default.
- W3126366999 sameAs 3126366999 @default.
- W3126366999 citedByCount "3" @default.
- W3126366999 countsByYear W31263669992021 @default.
- W3126366999 countsByYear W31263669992023 @default.
- W3126366999 crossrefType "journal-article" @default.
- W3126366999 hasAuthorship W3126366999A5074956841 @default.
- W3126366999 hasAuthorship W3126366999A5081147343 @default.
- W3126366999 hasBestOaLocation W31263669992 @default.
- W3126366999 hasConcept C108583219 @default.
- W3126366999 hasConcept C111472728 @default.
- W3126366999 hasConcept C119857082 @default.
- W3126366999 hasConcept C132010649 @default.
- W3126366999 hasConcept C138885662 @default.
- W3126366999 hasConcept C153180895 @default.
- W3126366999 hasConcept C154945302 @default.
- W3126366999 hasConcept C165696696 @default.
- W3126366999 hasConcept C202444582 @default.
- W3126366999 hasConcept C2776214188 @default.
- W3126366999 hasConcept C33923547 @default.
- W3126366999 hasConcept C38652104 @default.
- W3126366999 hasConcept C41008148 @default.
- W3126366999 hasConcept C45347329 @default.
- W3126366999 hasConcept C50644808 @default.
- W3126366999 hasConcept C89600930 @default.
- W3126366999 hasConcept C9652623 @default.
- W3126366999 hasConceptScore W3126366999C108583219 @default.
- W3126366999 hasConceptScore W3126366999C111472728 @default.
- W3126366999 hasConceptScore W3126366999C119857082 @default.
- W3126366999 hasConceptScore W3126366999C132010649 @default.
- W3126366999 hasConceptScore W3126366999C138885662 @default.
- W3126366999 hasConceptScore W3126366999C153180895 @default.
- W3126366999 hasConceptScore W3126366999C154945302 @default.
- W3126366999 hasConceptScore W3126366999C165696696 @default.
- W3126366999 hasConceptScore W3126366999C202444582 @default.
- W3126366999 hasConceptScore W3126366999C2776214188 @default.
- W3126366999 hasConceptScore W3126366999C33923547 @default.
- W3126366999 hasConceptScore W3126366999C38652104 @default.
- W3126366999 hasConceptScore W3126366999C41008148 @default.
- W3126366999 hasConceptScore W3126366999C45347329 @default.
- W3126366999 hasConceptScore W3126366999C50644808 @default.
- W3126366999 hasConceptScore W3126366999C89600930 @default.
- W3126366999 hasConceptScore W3126366999C9652623 @default.
- W3126366999 hasLocation W31263669991 @default.
- W3126366999 hasLocation W31263669992 @default.
- W3126366999 hasLocation W31263669993 @default.
- W3126366999 hasOpenAccess W3126366999 @default.
- W3126366999 hasPrimaryLocation W31263669991 @default.
- W3126366999 hasRelatedWork W2790662084 @default.
- W3126366999 hasRelatedWork W3014300295 @default.
- W3126366999 hasRelatedWork W3164822677 @default.
- W3126366999 hasRelatedWork W4223943233 @default.
- W3126366999 hasRelatedWork W4225161397 @default.
- W3126366999 hasRelatedWork W4312200629 @default.
- W3126366999 hasRelatedWork W4360585206 @default.
- W3126366999 hasRelatedWork W4364306694 @default.
- W3126366999 hasRelatedWork W4380075502 @default.
- W3126366999 hasRelatedWork W4380086463 @default.
- W3126366999 hasVolume "438" @default.
- W3126366999 isParatext "false" @default.
- W3126366999 isRetracted "false" @default.
- W3126366999 magId "3126366999" @default.
- W3126366999 workType "article" @default.