Matches in SemOpenAlex for { <https://semopenalex.org/work/W4304080274> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W4304080274 abstract "Image captioning is shown to be able to achieve a better performance by using scene graphs to represent the relations of objects in the image. The current captioning encoders generally use a Graph Convolutional Net (GCN) to represent the relation information and merge it with the object region features via concatenation or convolution to get the final input for sentence decoding. However, the GCN-based encoders in the existing methods are less effective for captioning due to two reasons. First, using the image captioning as the objective (i.e., Maximum Likelihood Estimation) rather than a relation-centric loss cannot fully explore the potential of the encoder. Second, using a pre-trained model instead of the encoder itself to extract the relationships is not flexible and cannot contribute to the explainability of the model. To improve the quality of image captioning, we propose a novel architecture ReFormer- a RElational transFORMER to generate features with relation information embedded and to explicitly express the pair-wise relationships between objects in the image. ReFormer incorporates the objective of scene graph generation with that of image captioning using one modified Transformer model. This design allows ReFormer to generate not only better image captions with the benefit of extracting strong relational image features, but also scene graphs to explicitly describe the pair-wise relationships. Experiments on publicly available datasets show that our model significantly outperforms state-of-the-art methods on image captioning and scene graph generation." @default.
- W4304080274 created "2022-10-10" @default.
- W4304080274 creator A5006822602 @default.
- W4304080274 creator A5036705554 @default.
- W4304080274 creator A5054023368 @default.
- W4304080274 date "2022-10-10" @default.
- W4304080274 modified "2023-10-11" @default.
- W4304080274 title "ReFormer: The Relational Transformer for Image Captioning" @default.
- W4304080274 cites W1956340063 @default.
- W4304080274 cites W2045284229 @default.
- W4304080274 cites W2133459682 @default.
- W4304080274 cites W2250539671 @default.
- W4304080274 cites W2481240925 @default.
- W4304080274 cites W2890531016 @default.
- W4304080274 cites W2939888942 @default.
- W4304080274 cites W2963260436 @default.
- W4304080274 cites W2963360627 @default.
- W4304080274 cites W2963565375 @default.
- W4304080274 cites W2963758027 @default.
- W4304080274 cites W2963938081 @default.
- W4304080274 cites W2989377923 @default.
- W4304080274 cites W2990818246 @default.
- W4304080274 cites W3016211260 @default.
- W4304080274 cites W3034316193 @default.
- W4304080274 cites W3034655362 @default.
- W4304080274 cites W3035017890 @default.
- W4304080274 cites W3107492437 @default.
- W4304080274 cites W3107848485 @default.
- W4304080274 cites W3110157234 @default.
- W4304080274 cites W4206675125 @default.
- W4304080274 doi "https://doi.org/10.1145/3503161.3548409" @default.
- W4304080274 hasPublicationYear "2022" @default.
- W4304080274 type Work @default.
- W4304080274 citedByCount "7" @default.
- W4304080274 countsByYear W43040802742023 @default.
- W4304080274 crossrefType "proceedings-article" @default.
- W4304080274 hasAuthorship W4304080274A5006822602 @default.
- W4304080274 hasAuthorship W4304080274A5036705554 @default.
- W4304080274 hasAuthorship W4304080274A5054023368 @default.
- W4304080274 hasBestOaLocation W43040802742 @default.
- W4304080274 hasConcept C111919701 @default.
- W4304080274 hasConcept C11413529 @default.
- W4304080274 hasConcept C115961682 @default.
- W4304080274 hasConcept C118505674 @default.
- W4304080274 hasConcept C121332964 @default.
- W4304080274 hasConcept C132525143 @default.
- W4304080274 hasConcept C154945302 @default.
- W4304080274 hasConcept C157657479 @default.
- W4304080274 hasConcept C165801399 @default.
- W4304080274 hasConcept C179372163 @default.
- W4304080274 hasConcept C204321447 @default.
- W4304080274 hasConcept C205711294 @default.
- W4304080274 hasConcept C2777530160 @default.
- W4304080274 hasConcept C31972630 @default.
- W4304080274 hasConcept C41008148 @default.
- W4304080274 hasConcept C57273362 @default.
- W4304080274 hasConcept C62520636 @default.
- W4304080274 hasConcept C66322947 @default.
- W4304080274 hasConcept C80444323 @default.
- W4304080274 hasConceptScore W4304080274C111919701 @default.
- W4304080274 hasConceptScore W4304080274C11413529 @default.
- W4304080274 hasConceptScore W4304080274C115961682 @default.
- W4304080274 hasConceptScore W4304080274C118505674 @default.
- W4304080274 hasConceptScore W4304080274C121332964 @default.
- W4304080274 hasConceptScore W4304080274C132525143 @default.
- W4304080274 hasConceptScore W4304080274C154945302 @default.
- W4304080274 hasConceptScore W4304080274C157657479 @default.
- W4304080274 hasConceptScore W4304080274C165801399 @default.
- W4304080274 hasConceptScore W4304080274C179372163 @default.
- W4304080274 hasConceptScore W4304080274C204321447 @default.
- W4304080274 hasConceptScore W4304080274C205711294 @default.
- W4304080274 hasConceptScore W4304080274C2777530160 @default.
- W4304080274 hasConceptScore W4304080274C31972630 @default.
- W4304080274 hasConceptScore W4304080274C41008148 @default.
- W4304080274 hasConceptScore W4304080274C57273362 @default.
- W4304080274 hasConceptScore W4304080274C62520636 @default.
- W4304080274 hasConceptScore W4304080274C66322947 @default.
- W4304080274 hasConceptScore W4304080274C80444323 @default.
- W4304080274 hasLocation W43040802741 @default.
- W4304080274 hasLocation W43040802742 @default.
- W4304080274 hasOpenAccess W4304080274 @default.
- W4304080274 hasPrimaryLocation W43040802741 @default.
- W4304080274 hasRelatedWork W2547835662 @default.
- W4304080274 hasRelatedWork W3025136821 @default.
- W4304080274 hasRelatedWork W3035237998 @default.
- W4304080274 hasRelatedWork W3093329502 @default.
- W4304080274 hasRelatedWork W4224046780 @default.
- W4304080274 hasRelatedWork W4281560470 @default.
- W4304080274 hasRelatedWork W4312545247 @default.
- W4304080274 hasRelatedWork W4312845724 @default.
- W4304080274 hasRelatedWork W4384210086 @default.
- W4304080274 hasRelatedWork W4385606240 @default.
- W4304080274 isParatext "false" @default.
- W4304080274 isRetracted "false" @default.
- W4304080274 workType "article" @default.