Matches in SemOpenAlex for { <https://semopenalex.org/work/W4205202009> ?p ?o ?g. }
- W4205202009 endingPage "4952" @default.
- W4205202009 startingPage "4940" @default.
- W4205202009 abstract "Recognition and localization models can be generally decomposed into three components: encoder, decoder, and task head. In this paper, we rethink the necessity of decoder, as we observe that it brings additional computational and parametric burden. We thus propose to remove the decoder and present a bidirectional cross-scale architecture that is able to obtain rich semantic information and precise localization in a unified backbone. Extensive experiments demonstrate that, different from common encoder-decoder models and other down-sampling and up-sampling backbones, the proposed BiCSNet achieves improved performances compared to existing architectures for pixel-level tasks. In object detection, our BiCSNet brings significant performance improvement by ~ 3% AP at various scales with 13% – 23% fewer FLOPS, compared with ResNet-FPN models on COCO dataset. In Instance segmentation, the AP can be improved by 1% over SpineNet. BiCSNet is also promising for semantic segmentation tasks, as the proposed BiCSNet pre-trained on ImageNet alone significantly outperforms DeepLabv3 pre-trained on both ImageNet and COCO dataset by 1.3% in mIOU with 89% fewer FLOPs on PASCAL VOC 2012." @default.
- W4205202009 created "2022-01-25" @default.
- W4205202009 creator A5009279384 @default.
- W4205202009 creator A5019191351 @default.
- W4205202009 creator A5030942549 @default.
- W4205202009 creator A5060706069 @default.
- W4205202009 creator A5068120645 @default.
- W4205202009 creator A5070629154 @default.
- W4205202009 date "2022-08-01" @default.
- W4205202009 modified "2023-09-27" @default.
- W4205202009 title "BiCSNet: A Bidirectional Cross-Scale Backbone for Recognition and Localization" @default.
- W4205202009 cites W2097117768 @default.
- W4205202009 cites W2102605133 @default.
- W4205202009 cites W2117539524 @default.
- W4205202009 cites W2144794286 @default.
- W4205202009 cites W2194775991 @default.
- W4205202009 cites W2307770531 @default.
- W4205202009 cites W2549139847 @default.
- W4205202009 cites W2557728737 @default.
- W4205202009 cites W2752782242 @default.
- W4205202009 cites W2797977484 @default.
- W4205202009 cites W2888728082 @default.
- W4205202009 cites W2916798096 @default.
- W4205202009 cites W2922509574 @default.
- W4205202009 cites W2963037989 @default.
- W4205202009 cites W2963125010 @default.
- W4205202009 cites W2963150697 @default.
- W4205202009 cites W2963163009 @default.
- W4205202009 cites W2963351448 @default.
- W4205202009 cites W2963849369 @default.
- W4205202009 cites W2963855133 @default.
- W4205202009 cites W2963857746 @default.
- W4205202009 cites W2964241181 @default.
- W4205202009 cites W2964444661 @default.
- W4205202009 cites W2972515477 @default.
- W4205202009 cites W2982083293 @default.
- W4205202009 cites W2984009799 @default.
- W4205202009 cites W2985228975 @default.
- W4205202009 cites W2989604896 @default.
- W4205202009 cites W3009977275 @default.
- W4205202009 cites W3034971973 @default.
- W4205202009 cites W3035087309 @default.
- W4205202009 cites W3035396860 @default.
- W4205202009 cites W3035658156 @default.
- W4205202009 cites W3042011474 @default.
- W4205202009 cites W3088979451 @default.
- W4205202009 cites W3096493942 @default.
- W4205202009 cites W3100179319 @default.
- W4205202009 cites W3100321043 @default.
- W4205202009 cites W3131500599 @default.
- W4205202009 cites W3138516171 @default.
- W4205202009 cites W3170952646 @default.
- W4205202009 cites W3179888767 @default.
- W4205202009 cites W3198011468 @default.
- W4205202009 cites W2993592664 @default.
- W4205202009 doi "https://doi.org/10.1109/tcsvt.2021.3138743" @default.
- W4205202009 hasPublicationYear "2022" @default.
- W4205202009 type Work @default.
- W4205202009 citedByCount "0" @default.
- W4205202009 crossrefType "journal-article" @default.
- W4205202009 hasAuthorship W4205202009A5009279384 @default.
- W4205202009 hasAuthorship W4205202009A5019191351 @default.
- W4205202009 hasAuthorship W4205202009A5030942549 @default.
- W4205202009 hasAuthorship W4205202009A5060706069 @default.
- W4205202009 hasAuthorship W4205202009A5068120645 @default.
- W4205202009 hasAuthorship W4205202009A5070629154 @default.
- W4205202009 hasConcept C105795698 @default.
- W4205202009 hasConcept C108583219 @default.
- W4205202009 hasConcept C111919701 @default.
- W4205202009 hasConcept C11413529 @default.
- W4205202009 hasConcept C117251300 @default.
- W4205202009 hasConcept C118505674 @default.
- W4205202009 hasConcept C153180895 @default.
- W4205202009 hasConcept C154945302 @default.
- W4205202009 hasConcept C173608175 @default.
- W4205202009 hasConcept C179799912 @default.
- W4205202009 hasConcept C199360897 @default.
- W4205202009 hasConcept C2776151529 @default.
- W4205202009 hasConcept C2944601119 @default.
- W4205202009 hasConcept C31258907 @default.
- W4205202009 hasConcept C33923547 @default.
- W4205202009 hasConcept C3826847 @default.
- W4205202009 hasConcept C41008148 @default.
- W4205202009 hasConcept C75608658 @default.
- W4205202009 hasConcept C88796919 @default.
- W4205202009 hasConcept C89600930 @default.
- W4205202009 hasConceptScore W4205202009C105795698 @default.
- W4205202009 hasConceptScore W4205202009C108583219 @default.
- W4205202009 hasConceptScore W4205202009C111919701 @default.
- W4205202009 hasConceptScore W4205202009C11413529 @default.
- W4205202009 hasConceptScore W4205202009C117251300 @default.
- W4205202009 hasConceptScore W4205202009C118505674 @default.
- W4205202009 hasConceptScore W4205202009C153180895 @default.
- W4205202009 hasConceptScore W4205202009C154945302 @default.
- W4205202009 hasConceptScore W4205202009C173608175 @default.
- W4205202009 hasConceptScore W4205202009C179799912 @default.
- W4205202009 hasConceptScore W4205202009C199360897 @default.
- W4205202009 hasConceptScore W4205202009C2776151529 @default.