Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312454920> ?p ?o ?g. }
- W4312454920 abstract "Deep convolutional neural networks have considerably improved state-of-the-art results for semantic segmentation. Nevertheless, even modern architectures lack the ability to generalize well to a test dataset that originates from a different domain. To avoid the costly annotation of training data for unseen domains, unsupervised domain adaptation (UDA) attempts to provide efficient knowledge transfer from a labeled source domain to an unlabeled target domain. Previous work has mainly focused on minimizing the discrepancy between the two domains by using adversarial training or self-training. While adversarial training may fail to align the correct semantic categories as it minimizes the discrepancy between the global distributions, self-training raises the question of how to provide reliable pseudolabels. To align the correct semantic categories across domains, we propose a contrastive learning approach that adapts category-wise centroids across domains. Furthermore, we extend our method with self-training, where we use a memory-efficient temporal ensemble to generate consistent and reliable pseudo-labels. Although both contrastive learning and self-training (CLST) through temporal ensembling enable knowledge transfer between two domains, it is their combination that leads to a symbiotic structure. We validate our approach on two domain adaptation benchmarks: GTA5 → Cityscapes and SYNTHIA → Cityscapes. Our method achieves better results than the state-of-the-art." @default.
- W4312454920 created "2023-01-04" @default.
- W4312454920 creator A5021977554 @default.
- W4312454920 creator A5046815454 @default.
- W4312454920 creator A5061072628 @default.
- W4312454920 creator A5090266268 @default.
- W4312454920 date "2022-07-18" @default.
- W4312454920 modified "2023-10-01" @default.
- W4312454920 title "Contrastive Learning and Self-Training for Unsupervised Domain Adaptation in Semantic Segmentation" @default.
- W4312454920 cites W1779010541 @default.
- W4312454920 cites W2138621090 @default.
- W4312454920 cites W2340897893 @default.
- W4312454920 cites W2412782625 @default.
- W4312454920 cites W2431874326 @default.
- W4312454920 cites W2487365028 @default.
- W4312454920 cites W2895281799 @default.
- W4312454920 cites W2905535599 @default.
- W4312454920 cites W2963073217 @default.
- W4312454920 cites W2963107255 @default.
- W4312454920 cites W2963118547 @default.
- W4312454920 cites W2963209464 @default.
- W4312454920 cites W2963532621 @default.
- W4312454920 cites W2964285681 @default.
- W4312454920 cites W2964309882 @default.
- W4312454920 cites W2969893028 @default.
- W4312454920 cites W2972285644 @default.
- W4312454920 cites W2981429991 @default.
- W4312454920 cites W2981512393 @default.
- W4312454920 cites W2985406498 @default.
- W4312454920 cites W2987385519 @default.
- W4312454920 cites W3034417116 @default.
- W4312454920 cites W3035294798 @default.
- W4312454920 cites W3035524453 @default.
- W4312454920 cites W3102977943 @default.
- W4312454920 cites W3108566666 @default.
- W4312454920 cites W3119635706 @default.
- W4312454920 cites W3120804725 @default.
- W4312454920 cites W3122412340 @default.
- W4312454920 cites W3157653192 @default.
- W4312454920 doi "https://doi.org/10.1109/ijcnn55064.2022.9892322" @default.
- W4312454920 hasPublicationYear "2022" @default.
- W4312454920 type Work @default.
- W4312454920 citedByCount "6" @default.
- W4312454920 countsByYear W43124549202021 @default.
- W4312454920 countsByYear W43124549202022 @default.
- W4312454920 countsByYear W43124549202023 @default.
- W4312454920 crossrefType "proceedings-article" @default.
- W4312454920 hasAuthorship W4312454920A5021977554 @default.
- W4312454920 hasAuthorship W4312454920A5046815454 @default.
- W4312454920 hasAuthorship W4312454920A5061072628 @default.
- W4312454920 hasAuthorship W4312454920A5090266268 @default.
- W4312454920 hasBestOaLocation W43124549202 @default.
- W4312454920 hasConcept C108583219 @default.
- W4312454920 hasConcept C119857082 @default.
- W4312454920 hasConcept C120665830 @default.
- W4312454920 hasConcept C121332964 @default.
- W4312454920 hasConcept C134306372 @default.
- W4312454920 hasConcept C139807058 @default.
- W4312454920 hasConcept C146599234 @default.
- W4312454920 hasConcept C150899416 @default.
- W4312454920 hasConcept C153180895 @default.
- W4312454920 hasConcept C154945302 @default.
- W4312454920 hasConcept C184337299 @default.
- W4312454920 hasConcept C199360897 @default.
- W4312454920 hasConcept C204321447 @default.
- W4312454920 hasConcept C33923547 @default.
- W4312454920 hasConcept C36503486 @default.
- W4312454920 hasConcept C37736160 @default.
- W4312454920 hasConcept C41008148 @default.
- W4312454920 hasConcept C81363708 @default.
- W4312454920 hasConcept C89600930 @default.
- W4312454920 hasConceptScore W4312454920C108583219 @default.
- W4312454920 hasConceptScore W4312454920C119857082 @default.
- W4312454920 hasConceptScore W4312454920C120665830 @default.
- W4312454920 hasConceptScore W4312454920C121332964 @default.
- W4312454920 hasConceptScore W4312454920C134306372 @default.
- W4312454920 hasConceptScore W4312454920C139807058 @default.
- W4312454920 hasConceptScore W4312454920C146599234 @default.
- W4312454920 hasConceptScore W4312454920C150899416 @default.
- W4312454920 hasConceptScore W4312454920C153180895 @default.
- W4312454920 hasConceptScore W4312454920C154945302 @default.
- W4312454920 hasConceptScore W4312454920C184337299 @default.
- W4312454920 hasConceptScore W4312454920C199360897 @default.
- W4312454920 hasConceptScore W4312454920C204321447 @default.
- W4312454920 hasConceptScore W4312454920C33923547 @default.
- W4312454920 hasConceptScore W4312454920C36503486 @default.
- W4312454920 hasConceptScore W4312454920C37736160 @default.
- W4312454920 hasConceptScore W4312454920C41008148 @default.
- W4312454920 hasConceptScore W4312454920C81363708 @default.
- W4312454920 hasConceptScore W4312454920C89600930 @default.
- W4312454920 hasFunder F4320323803 @default.
- W4312454920 hasLocation W43124549201 @default.
- W4312454920 hasLocation W43124549202 @default.
- W4312454920 hasOpenAccess W4312454920 @default.
- W4312454920 hasPrimaryLocation W43124549201 @default.
- W4312454920 hasRelatedWork W2996856019 @default.
- W4312454920 hasRelatedWork W3012459282 @default.
- W4312454920 hasRelatedWork W3018421652 @default.
- W4312454920 hasRelatedWork W3021430260 @default.
- W4312454920 hasRelatedWork W3091976719 @default.