Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386071498> ?p ?o ?g. }
Showing items 1 to 98 of
98
with 100 items per page.
- W4386071498 abstract "Current state-of-the-art image-text matching methods implicitly align the visual-semantic fragments, like regions in images and words in sentences, and adopt cross-attention mechanism to discover fine-grained cross-modal semantic correspondence. However, the cross-attention mechanism may bring redundant or irrelevant region-word alignments, degenerating retrieval accuracy and limiting efficiency. Although many researchers have made progress in mining meaningful alignments and thus improving accuracy, the problem of poor efficiency remains unresolved. In this work, we propose to learn fine-grained image-text matching from the perspective of information coding. Specifically, we suggest a coding framework to explain the fragments aligning process, which provides a novel view to reexamine the cross-attention mechanism and analyze the problem of redundant alignments. Based on this framework, a Cross-modal Hard Aligning Network (CHAN) is designed, which comprehensively exploits the most relevant region-word pairs and eliminates all other alignments. Extensive experiments conducted on two public datasets, MS-COCO and Flickr30K, verify that the relevance of the most associated word-region pairs is discriminative enough as an indicator of the image-text similarity, with superior accuracy and efficiency over the state-of-the-art approaches on the bidirectional image and text retrieval tasks. Our code will be available at https://github.com/ppanzx/CHAN." @default.
- W4386071498 created "2023-08-23" @default.
- W4386071498 creator A5005961599 @default.
- W4386071498 creator A5026986847 @default.
- W4386071498 creator A5086628680 @default.
- W4386071498 date "2023-06-01" @default.
- W4386071498 modified "2023-09-27" @default.
- W4386071498 title "Fine-grained Image-text Matching by Cross-modal Hard Aligning Network" @default.
- W4386071498 cites W1905882502 @default.
- W4386071498 cites W2002273623 @default.
- W4386071498 cites W2131846894 @default.
- W4386071498 cites W2145406111 @default.
- W4386071498 cites W2155490028 @default.
- W4386071498 cites W2185175083 @default.
- W4386071498 cites W2194775991 @default.
- W4386071498 cites W2250539671 @default.
- W4386071498 cites W2277195237 @default.
- W4386071498 cites W2552579943 @default.
- W4386071498 cites W2606473278 @default.
- W4386071498 cites W2745461083 @default.
- W4386071498 cites W2765440071 @default.
- W4386071498 cites W2962964995 @default.
- W4386071498 cites W2981586349 @default.
- W4386071498 cites W2982078236 @default.
- W4386071498 cites W2988823324 @default.
- W4386071498 cites W3035454331 @default.
- W4386071498 cites W3035588244 @default.
- W4386071498 cites W3035605030 @default.
- W4386071498 cites W3036625283 @default.
- W4386071498 cites W3171345413 @default.
- W4386071498 cites W3175888430 @default.
- W4386071498 cites W3213100861 @default.
- W4386071498 cites W4214819138 @default.
- W4386071498 cites W4312761738 @default.
- W4386071498 cites W639708223 @default.
- W4386071498 doi "https://doi.org/10.1109/cvpr52729.2023.01847" @default.
- W4386071498 hasPublicationYear "2023" @default.
- W4386071498 type Work @default.
- W4386071498 citedByCount "0" @default.
- W4386071498 crossrefType "proceedings-article" @default.
- W4386071498 hasAuthorship W4386071498A5005961599 @default.
- W4386071498 hasAuthorship W4386071498A5026986847 @default.
- W4386071498 hasAuthorship W4386071498A5086628680 @default.
- W4386071498 hasConcept C105795698 @default.
- W4386071498 hasConcept C138885662 @default.
- W4386071498 hasConcept C153180895 @default.
- W4386071498 hasConcept C154945302 @default.
- W4386071498 hasConcept C158154518 @default.
- W4386071498 hasConcept C165064840 @default.
- W4386071498 hasConcept C17744445 @default.
- W4386071498 hasConcept C179518139 @default.
- W4386071498 hasConcept C185592680 @default.
- W4386071498 hasConcept C188027245 @default.
- W4386071498 hasConcept C199539241 @default.
- W4386071498 hasConcept C204321447 @default.
- W4386071498 hasConcept C23123220 @default.
- W4386071498 hasConcept C33923547 @default.
- W4386071498 hasConcept C41008148 @default.
- W4386071498 hasConcept C41895202 @default.
- W4386071498 hasConcept C71139939 @default.
- W4386071498 hasConcept C90805587 @default.
- W4386071498 hasConcept C97931131 @default.
- W4386071498 hasConceptScore W4386071498C105795698 @default.
- W4386071498 hasConceptScore W4386071498C138885662 @default.
- W4386071498 hasConceptScore W4386071498C153180895 @default.
- W4386071498 hasConceptScore W4386071498C154945302 @default.
- W4386071498 hasConceptScore W4386071498C158154518 @default.
- W4386071498 hasConceptScore W4386071498C165064840 @default.
- W4386071498 hasConceptScore W4386071498C17744445 @default.
- W4386071498 hasConceptScore W4386071498C179518139 @default.
- W4386071498 hasConceptScore W4386071498C185592680 @default.
- W4386071498 hasConceptScore W4386071498C188027245 @default.
- W4386071498 hasConceptScore W4386071498C199539241 @default.
- W4386071498 hasConceptScore W4386071498C204321447 @default.
- W4386071498 hasConceptScore W4386071498C23123220 @default.
- W4386071498 hasConceptScore W4386071498C33923547 @default.
- W4386071498 hasConceptScore W4386071498C41008148 @default.
- W4386071498 hasConceptScore W4386071498C41895202 @default.
- W4386071498 hasConceptScore W4386071498C71139939 @default.
- W4386071498 hasConceptScore W4386071498C90805587 @default.
- W4386071498 hasConceptScore W4386071498C97931131 @default.
- W4386071498 hasFunder F4320321001 @default.
- W4386071498 hasLocation W43860714981 @default.
- W4386071498 hasOpenAccess W4386071498 @default.
- W4386071498 hasPrimaryLocation W43860714981 @default.
- W4386071498 hasRelatedWork W1972656095 @default.
- W4386071498 hasRelatedWork W2024160000 @default.
- W4386071498 hasRelatedWork W2061273563 @default.
- W4386071498 hasRelatedWork W2285052147 @default.
- W4386071498 hasRelatedWork W2729514902 @default.
- W4386071498 hasRelatedWork W2743258233 @default.
- W4386071498 hasRelatedWork W2773500201 @default.
- W4386071498 hasRelatedWork W2970216048 @default.
- W4386071498 hasRelatedWork W2998168123 @default.
- W4386071498 hasRelatedWork W4287995534 @default.
- W4386071498 isParatext "false" @default.
- W4386071498 isRetracted "false" @default.
- W4386071498 workType "article" @default.