Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312543911> ?p ?o ?g. }
- W4312543911 abstract "Referring image segmentation is an advanced semantic segmentation task where target is not a predefined class but is described in natural language. Most of existing methods for this task rely heavily on convolutional neural networks, which however have trouble capturing long-range dependencies between entities in the language expression and are not flexible enough for modeling interactions between the two different modalities. To address these issues, we present the first convolution-free model for referring image segmentation using transformers, dubbed ReSTR. Since it extracts features of both modalities through transformer encoders, it can capture long-range dependencies between entities within each modality. Also, ReSTR fuses features of the two modalities by a self-attention encoder, which enables flexible and adaptive interactions between the two modalities in the fusion process. The fused features are fed to a segmentation module, which works adaptively according to the image and language expression in hand. ReSTR is evaluated and compared with previous work on all public benchmarks, where it outperforms all existing models." @default.
- W4312543911 created "2023-01-05" @default.
- W4312543911 creator A5044840923 @default.
- W4312543911 creator A5049963367 @default.
- W4312543911 creator A5051469345 @default.
- W4312543911 creator A5060343759 @default.
- W4312543911 creator A5074668738 @default.
- W4312543911 date "2022-06-01" @default.
- W4312543911 modified "2023-10-16" @default.
- W4312543911 title "ReSTR: Convolution-free Referring Image Segmentation Using Transformers" @default.
- W4312543911 cites W1745334888 @default.
- W4312543911 cites W1903029394 @default.
- W4312543911 cites W2006147162 @default.
- W4312543911 cites W2108598243 @default.
- W4312543911 cites W2124592697 @default.
- W4312543911 cites W2250539671 @default.
- W4312543911 cites W2251512949 @default.
- W4312543911 cites W2560023338 @default.
- W4312543911 cites W2563705555 @default.
- W4312543911 cites W2605127024 @default.
- W4312543911 cites W2798556392 @default.
- W4312543911 cites W2799213142 @default.
- W4312543911 cites W2963091558 @default.
- W4312543911 cites W2963108253 @default.
- W4312543911 cites W2980088508 @default.
- W4312543911 cites W2981689412 @default.
- W4312543911 cites W2990775046 @default.
- W4312543911 cites W3004019157 @default.
- W4312543911 cites W3034325957 @default.
- W4312543911 cites W3034885317 @default.
- W4312543911 cites W3131500599 @default.
- W4312543911 cites W3138516171 @default.
- W4312543911 cites W3159619744 @default.
- W4312543911 cites W3168649818 @default.
- W4312543911 cites W3169998662 @default.
- W4312543911 cites W3172509117 @default.
- W4312543911 cites W3172522282 @default.
- W4312543911 cites W3187664142 @default.
- W4312543911 cites W3201770677 @default.
- W4312543911 cites W4214893857 @default.
- W4312543911 cites W4243246777 @default.
- W4312543911 doi "https://doi.org/10.1109/cvpr52688.2022.01761" @default.
- W4312543911 hasPublicationYear "2022" @default.
- W4312543911 type Work @default.
- W4312543911 citedByCount "18" @default.
- W4312543911 countsByYear W43125439112023 @default.
- W4312543911 crossrefType "proceedings-article" @default.
- W4312543911 hasAuthorship W4312543911A5044840923 @default.
- W4312543911 hasAuthorship W4312543911A5049963367 @default.
- W4312543911 hasAuthorship W4312543911A5051469345 @default.
- W4312543911 hasAuthorship W4312543911A5060343759 @default.
- W4312543911 hasAuthorship W4312543911A5074668738 @default.
- W4312543911 hasBestOaLocation W43125439112 @default.
- W4312543911 hasConcept C111919701 @default.
- W4312543911 hasConcept C118505674 @default.
- W4312543911 hasConcept C119599485 @default.
- W4312543911 hasConcept C124504099 @default.
- W4312543911 hasConcept C127413603 @default.
- W4312543911 hasConcept C144024400 @default.
- W4312543911 hasConcept C153180895 @default.
- W4312543911 hasConcept C154945302 @default.
- W4312543911 hasConcept C165801399 @default.
- W4312543911 hasConcept C2779903281 @default.
- W4312543911 hasConcept C31972630 @default.
- W4312543911 hasConcept C36289849 @default.
- W4312543911 hasConcept C41008148 @default.
- W4312543911 hasConcept C66322947 @default.
- W4312543911 hasConcept C81363708 @default.
- W4312543911 hasConcept C89600930 @default.
- W4312543911 hasConceptScore W4312543911C111919701 @default.
- W4312543911 hasConceptScore W4312543911C118505674 @default.
- W4312543911 hasConceptScore W4312543911C119599485 @default.
- W4312543911 hasConceptScore W4312543911C124504099 @default.
- W4312543911 hasConceptScore W4312543911C127413603 @default.
- W4312543911 hasConceptScore W4312543911C144024400 @default.
- W4312543911 hasConceptScore W4312543911C153180895 @default.
- W4312543911 hasConceptScore W4312543911C154945302 @default.
- W4312543911 hasConceptScore W4312543911C165801399 @default.
- W4312543911 hasConceptScore W4312543911C2779903281 @default.
- W4312543911 hasConceptScore W4312543911C31972630 @default.
- W4312543911 hasConceptScore W4312543911C36289849 @default.
- W4312543911 hasConceptScore W4312543911C41008148 @default.
- W4312543911 hasConceptScore W4312543911C66322947 @default.
- W4312543911 hasConceptScore W4312543911C81363708 @default.
- W4312543911 hasConceptScore W4312543911C89600930 @default.
- W4312543911 hasFunder F4320320671 @default.
- W4312543911 hasFunder F4320335489 @default.
- W4312543911 hasLocation W43125439111 @default.
- W4312543911 hasLocation W43125439112 @default.
- W4312543911 hasOpenAccess W4312543911 @default.
- W4312543911 hasPrimaryLocation W43125439111 @default.
- W4312543911 hasRelatedWork W1669643531 @default.
- W4312543911 hasRelatedWork W1982826852 @default.
- W4312543911 hasRelatedWork W2005437358 @default.
- W4312543911 hasRelatedWork W2008656436 @default.
- W4312543911 hasRelatedWork W2023558673 @default.
- W4312543911 hasRelatedWork W2110230079 @default.
- W4312543911 hasRelatedWork W2134924024 @default.
- W4312543911 hasRelatedWork W2517104666 @default.
- W4312543911 hasRelatedWork W2613186388 @default.