Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313591097> ?p ?o ?g. }
- W4313591097 endingPage "16" @default.
- W4313591097 startingPage "1" @default.
- W4313591097 abstract "Recent advances in self-supervised learning (SSL) in computer vision are primarily comparative, whose goal is to preserve invariant and discriminative semantics in latent representations by comparing siamese image views. However, the preserved high-level semantics do not contain enough local information, which is vital in medical image analysis (e.g., image-based diagnosis and tumor segmentation). To mitigate the locality problem of comparative SSL, we propose to incorporate the task of pixel restoration for explicitly encoding more pixel-level information into high-level semantics. We also address the preservation of scale information, a powerful tool in aiding image understanding but has not drawn much attention in SSL. The resulting framework can be formulated as a multi-task optimization problem on the feature pyramid. Specifically, we conduct multi-scale pixel restoration and siamese feature comparison in the pyramid. In addition, we propose non-skip U-Net to build the feature pyramid and develop sub-crop to replace multi-crop in 3D medical imaging. The proposed unified SSL framework (PCRLv2) surpasses its self-supervised counterparts on various tasks, including brain tumor segmentation (BraTS 2018), chest pathology identification (ChestX-ray, CheXpert), pulmonary nodule detection (LUNA), and abdominal organ segmentation (LiTS), sometimes outperforming them by large margins with limited annotations. Codes and models are available at https://github.com/RL4M/PCRLv2." @default.
- W4313591097 created "2023-01-06" @default.
- W4313591097 creator A5006431737 @default.
- W4313591097 creator A5012166205 @default.
- W4313591097 creator A5037799666 @default.
- W4313591097 creator A5071959906 @default.
- W4313591097 creator A5072781957 @default.
- W4313591097 date "2023-01-01" @default.
- W4313591097 modified "2023-10-01" @default.
- W4313591097 title "A Unified Visual Information Preservation Framework for Self-supervised Pre-training in Medical Image Analysis" @default.
- W4313591097 cites W1903029394 @default.
- W4313591097 cites W1986649315 @default.
- W4313591097 cites W2108598243 @default.
- W4313591097 cites W2151103935 @default.
- W4313591097 cites W2161969291 @default.
- W4313591097 cites W2194775991 @default.
- W4313591097 cites W2295107390 @default.
- W4313591097 cites W2308529009 @default.
- W4313591097 cites W2558661413 @default.
- W4313591097 cites W2565639579 @default.
- W4313591097 cites W2584017349 @default.
- W4313591097 cites W2963420272 @default.
- W4313591097 cites W2963466845 @default.
- W4313591097 cites W2964744899 @default.
- W4313591097 cites W300523764 @default.
- W4313591097 cites W3012412627 @default.
- W4313591097 cites W3016836174 @default.
- W4313591097 cites W3035524453 @default.
- W4313591097 cites W3092603779 @default.
- W4313591097 cites W3093130481 @default.
- W4313591097 cites W3101156210 @default.
- W4313591097 cites W3108655343 @default.
- W4313591097 cites W3112701542 @default.
- W4313591097 cites W3120430728 @default.
- W4313591097 cites W3129875423 @default.
- W4313591097 cites W3171007011 @default.
- W4313591097 cites W3171849353 @default.
- W4313591097 cites W3174535521 @default.
- W4313591097 cites W3194662286 @default.
- W4313591097 cites W3196792593 @default.
- W4313591097 cites W3203378019 @default.
- W4313591097 doi "https://doi.org/10.1109/tpami.2023.3234002" @default.
- W4313591097 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37018263" @default.
- W4313591097 hasPublicationYear "2023" @default.
- W4313591097 type Work @default.
- W4313591097 citedByCount "3" @default.
- W4313591097 countsByYear W43135910972023 @default.
- W4313591097 crossrefType "journal-article" @default.
- W4313591097 hasAuthorship W4313591097A5006431737 @default.
- W4313591097 hasAuthorship W4313591097A5012166205 @default.
- W4313591097 hasAuthorship W4313591097A5037799666 @default.
- W4313591097 hasAuthorship W4313591097A5071959906 @default.
- W4313591097 hasAuthorship W4313591097A5072781957 @default.
- W4313591097 hasConcept C119857082 @default.
- W4313591097 hasConcept C120665830 @default.
- W4313591097 hasConcept C121332964 @default.
- W4313591097 hasConcept C124504099 @default.
- W4313591097 hasConcept C138885662 @default.
- W4313591097 hasConcept C142575187 @default.
- W4313591097 hasConcept C153180895 @default.
- W4313591097 hasConcept C154945302 @default.
- W4313591097 hasConcept C160633673 @default.
- W4313591097 hasConcept C184337299 @default.
- W4313591097 hasConcept C199360897 @default.
- W4313591097 hasConcept C2776401178 @default.
- W4313591097 hasConcept C31601959 @default.
- W4313591097 hasConcept C31972630 @default.
- W4313591097 hasConcept C41008148 @default.
- W4313591097 hasConcept C41895202 @default.
- W4313591097 hasConcept C52622490 @default.
- W4313591097 hasConcept C89600930 @default.
- W4313591097 hasConcept C97931131 @default.
- W4313591097 hasConceptScore W4313591097C119857082 @default.
- W4313591097 hasConceptScore W4313591097C120665830 @default.
- W4313591097 hasConceptScore W4313591097C121332964 @default.
- W4313591097 hasConceptScore W4313591097C124504099 @default.
- W4313591097 hasConceptScore W4313591097C138885662 @default.
- W4313591097 hasConceptScore W4313591097C142575187 @default.
- W4313591097 hasConceptScore W4313591097C153180895 @default.
- W4313591097 hasConceptScore W4313591097C154945302 @default.
- W4313591097 hasConceptScore W4313591097C160633673 @default.
- W4313591097 hasConceptScore W4313591097C184337299 @default.
- W4313591097 hasConceptScore W4313591097C199360897 @default.
- W4313591097 hasConceptScore W4313591097C2776401178 @default.
- W4313591097 hasConceptScore W4313591097C31601959 @default.
- W4313591097 hasConceptScore W4313591097C31972630 @default.
- W4313591097 hasConceptScore W4313591097C41008148 @default.
- W4313591097 hasConceptScore W4313591097C41895202 @default.
- W4313591097 hasConceptScore W4313591097C52622490 @default.
- W4313591097 hasConceptScore W4313591097C89600930 @default.
- W4313591097 hasConceptScore W4313591097C97931131 @default.
- W4313591097 hasLocation W43135910971 @default.
- W4313591097 hasLocation W43135910972 @default.
- W4313591097 hasOpenAccess W4313591097 @default.
- W4313591097 hasPrimaryLocation W43135910971 @default.
- W4313591097 hasRelatedWork W2024160000 @default.
- W4313591097 hasRelatedWork W2136485282 @default.
- W4313591097 hasRelatedWork W2285052147 @default.