Matches in SemOpenAlex for { <https://semopenalex.org/work/W4304098587> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W4304098587 abstract "Recently, it is gaining increasingly attention to incorporate self-supervised technologies into few-shot learning. Previous methods have exclusively focused on image-level self-supervision, but they ignore that capturing subtle part features plays an important role in distinguishing fine-grained images. In this paper, we propose an approach named PaCL that embeds part-level contrastive learning into fine-grained few-shot image classification, strengthening the models' capability to extract discriminative features from indistinguishable images. PaCL treats parts as the inputs of contrastive learning, and it uses a transformation module to involve image-specific information into pre-defined meta parts, generating multiple features from each meta part depending on different images. To alleviate the impact of changes in views or occlusions, we propose to adopt part prototypes in contrastive learning. Part prototypes are generated by aggregating the features of each certain type of part, which are more reliable than directly using part features. A few-shot classifier is adopted to predict query images, which calculates the classification loss to optimize the transformation module and meta parts in conjunction with the loss calculated in contrastive learning. The optimization process will enforce the model to learn to extract discriminative and diverse features from different parts of the objects, even for the samples of unseen classes. Extensive studies show that our proposed method improves the performance of fine-grained few-shot image classification across several backbones, datasets, and tasks, achieving superior results compared with state-of-the-art methods." @default.
- W4304098587 created "2022-10-10" @default.
- W4304098587 creator A5044945447 @default.
- W4304098587 creator A5048944835 @default.
- W4304098587 creator A5054452935 @default.
- W4304098587 date "2022-10-10" @default.
- W4304098587 modified "2023-10-17" @default.
- W4304098587 title "PaCL: Part-level Contrastive Learning for Fine-grained Few-shot Image Classification" @default.
- W4304098587 cites W2138011018 @default.
- W4304098587 cites W2957105406 @default.
- W4304098587 cites W2965572487 @default.
- W4304098587 cites W3012255272 @default.
- W4304098587 cites W3034312118 @default.
- W4304098587 cites W3087963932 @default.
- W4304098587 cites W3107763055 @default.
- W4304098587 cites W3108975329 @default.
- W4304098587 cites W3128632573 @default.
- W4304098587 cites W3161622534 @default.
- W4304098587 cites W4206910437 @default.
- W4304098587 cites W4280538209 @default.
- W4304098587 doi "https://doi.org/10.1145/3503161.3547997" @default.
- W4304098587 hasPublicationYear "2022" @default.
- W4304098587 type Work @default.
- W4304098587 citedByCount "1" @default.
- W4304098587 countsByYear W43040985872023 @default.
- W4304098587 crossrefType "proceedings-article" @default.
- W4304098587 hasAuthorship W4304098587A5044945447 @default.
- W4304098587 hasAuthorship W4304098587A5048944835 @default.
- W4304098587 hasAuthorship W4304098587A5054452935 @default.
- W4304098587 hasConcept C104317684 @default.
- W4304098587 hasConcept C111919701 @default.
- W4304098587 hasConcept C115961682 @default.
- W4304098587 hasConcept C119857082 @default.
- W4304098587 hasConcept C153180895 @default.
- W4304098587 hasConcept C154945302 @default.
- W4304098587 hasConcept C185592680 @default.
- W4304098587 hasConcept C204241405 @default.
- W4304098587 hasConcept C41008148 @default.
- W4304098587 hasConcept C55493867 @default.
- W4304098587 hasConcept C75294576 @default.
- W4304098587 hasConcept C95623464 @default.
- W4304098587 hasConcept C97931131 @default.
- W4304098587 hasConcept C98045186 @default.
- W4304098587 hasConceptScore W4304098587C104317684 @default.
- W4304098587 hasConceptScore W4304098587C111919701 @default.
- W4304098587 hasConceptScore W4304098587C115961682 @default.
- W4304098587 hasConceptScore W4304098587C119857082 @default.
- W4304098587 hasConceptScore W4304098587C153180895 @default.
- W4304098587 hasConceptScore W4304098587C154945302 @default.
- W4304098587 hasConceptScore W4304098587C185592680 @default.
- W4304098587 hasConceptScore W4304098587C204241405 @default.
- W4304098587 hasConceptScore W4304098587C41008148 @default.
- W4304098587 hasConceptScore W4304098587C55493867 @default.
- W4304098587 hasConceptScore W4304098587C75294576 @default.
- W4304098587 hasConceptScore W4304098587C95623464 @default.
- W4304098587 hasConceptScore W4304098587C97931131 @default.
- W4304098587 hasConceptScore W4304098587C98045186 @default.
- W4304098587 hasLocation W43040985871 @default.
- W4304098587 hasOpenAccess W4304098587 @default.
- W4304098587 hasPrimaryLocation W43040985871 @default.
- W4304098587 hasRelatedWork W1997235926 @default.
- W4304098587 hasRelatedWork W2005234362 @default.
- W4304098587 hasRelatedWork W2153315159 @default.
- W4304098587 hasRelatedWork W2162970382 @default.
- W4304098587 hasRelatedWork W259157601 @default.
- W4304098587 hasRelatedWork W2761785940 @default.
- W4304098587 hasRelatedWork W2965546495 @default.
- W4304098587 hasRelatedWork W2971377935 @default.
- W4304098587 hasRelatedWork W3103844505 @default.
- W4304098587 hasRelatedWork W4205463238 @default.
- W4304098587 isParatext "false" @default.
- W4304098587 isRetracted "false" @default.
- W4304098587 workType "article" @default.