Matches in SemOpenAlex for { <https://semopenalex.org/work/W4328101918> ?p ?o ?g. }
Showing items 1 to 74 of
74
with 100 items per page.
- W4328101918 endingPage "109547" @default.
- W4328101918 startingPage "109547" @default.
- W4328101918 abstract "The task of fine-grained visual classification (FGVC) is to distinguish targets from subordinate classifications. Since fine-grained images have the inherent characteristic of large inter-class variances and small intra-class variances, it is considered an extremely difficult task. Most existing approaches adopt CNN-based networks as feature extractors, which causes the extracted discriminative regions to contain most parts of the object in this way, thus failing to locate the really important parts. Recently, the vision transformer (ViT) has demonstrated its power on a wide range of image tasks, which uses an attention mechanism to capture global contextual information to establish a remote dependency on the target and thus extract more powerful features. Nevertheless, the ViT model still focuses more on global coarse-grained information rather than local fine-grained information, which may lead to its undesirable performance in fine-grained image classification. To this end, we redesigned an attention aggregating transformer (AA-Trans) to better capture minor differences among images by improving the ViT structure in this paper. In detail, we propose a core attention aggregator (CAA), which enables better information sharing between each transformer layer. Besides, we further propose an innovative information entropy selector (IES) to guide the network in acquiring discriminative parts of the image precisely. Extensive experiments show that our proposed model structure can achieve a new state-of-the-art performance on several mainstream datasets." @default.
- W4328101918 created "2023-03-22" @default.
- W4328101918 creator A5006168425 @default.
- W4328101918 creator A5013288442 @default.
- W4328101918 creator A5015195367 @default.
- W4328101918 creator A5043308642 @default.
- W4328101918 creator A5060428189 @default.
- W4328101918 creator A5088303343 @default.
- W4328101918 date "2023-08-01" @default.
- W4328101918 modified "2023-09-27" @default.
- W4328101918 title "AA-trans: Core attention aggregating transformer with information entropy selector for fine-grained visual classification" @default.
- W4328101918 cites W3019723613 @default.
- W4328101918 cites W3040248848 @default.
- W4328101918 cites W3113809646 @default.
- W4328101918 cites W3124951096 @default.
- W4328101918 cites W3207898576 @default.
- W4328101918 cites W4220722342 @default.
- W4328101918 cites W4223574620 @default.
- W4328101918 cites W4224440221 @default.
- W4328101918 cites W4225307160 @default.
- W4328101918 cites W4280538209 @default.
- W4328101918 cites W4283361901 @default.
- W4328101918 doi "https://doi.org/10.1016/j.patcog.2023.109547" @default.
- W4328101918 hasPublicationYear "2023" @default.
- W4328101918 type Work @default.
- W4328101918 citedByCount "2" @default.
- W4328101918 countsByYear W43281019182023 @default.
- W4328101918 crossrefType "journal-article" @default.
- W4328101918 hasAuthorship W4328101918A5006168425 @default.
- W4328101918 hasAuthorship W4328101918A5013288442 @default.
- W4328101918 hasAuthorship W4328101918A5015195367 @default.
- W4328101918 hasAuthorship W4328101918A5043308642 @default.
- W4328101918 hasAuthorship W4328101918A5060428189 @default.
- W4328101918 hasAuthorship W4328101918A5088303343 @default.
- W4328101918 hasConcept C106301342 @default.
- W4328101918 hasConcept C119857082 @default.
- W4328101918 hasConcept C121332964 @default.
- W4328101918 hasConcept C153180895 @default.
- W4328101918 hasConcept C154945302 @default.
- W4328101918 hasConcept C165801399 @default.
- W4328101918 hasConcept C41008148 @default.
- W4328101918 hasConcept C62520636 @default.
- W4328101918 hasConcept C66322947 @default.
- W4328101918 hasConcept C97931131 @default.
- W4328101918 hasConceptScore W4328101918C106301342 @default.
- W4328101918 hasConceptScore W4328101918C119857082 @default.
- W4328101918 hasConceptScore W4328101918C121332964 @default.
- W4328101918 hasConceptScore W4328101918C153180895 @default.
- W4328101918 hasConceptScore W4328101918C154945302 @default.
- W4328101918 hasConceptScore W4328101918C165801399 @default.
- W4328101918 hasConceptScore W4328101918C41008148 @default.
- W4328101918 hasConceptScore W4328101918C62520636 @default.
- W4328101918 hasConceptScore W4328101918C66322947 @default.
- W4328101918 hasConceptScore W4328101918C97931131 @default.
- W4328101918 hasFunder F4320321927 @default.
- W4328101918 hasLocation W43281019181 @default.
- W4328101918 hasOpenAccess W4328101918 @default.
- W4328101918 hasPrimaryLocation W43281019181 @default.
- W4328101918 hasRelatedWork W1972656095 @default.
- W4328101918 hasRelatedWork W2024160000 @default.
- W4328101918 hasRelatedWork W2061273563 @default.
- W4328101918 hasRelatedWork W2285052147 @default.
- W4328101918 hasRelatedWork W2729514902 @default.
- W4328101918 hasRelatedWork W2743258233 @default.
- W4328101918 hasRelatedWork W2773500201 @default.
- W4328101918 hasRelatedWork W2970216048 @default.
- W4328101918 hasRelatedWork W2998168123 @default.
- W4328101918 hasRelatedWork W4287995534 @default.
- W4328101918 hasVolume "140" @default.
- W4328101918 isParatext "false" @default.
- W4328101918 isRetracted "false" @default.
- W4328101918 workType "article" @default.