Matches in SemOpenAlex for { <https://semopenalex.org/work/W4320518829> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4320518829 abstract "Computed tomography (CT) samples with pathological annotations are difficult to obtain. As a result, the computer-aided diagnosis (CAD) algorithms are trained on small datasets (e.g., LIDC-IDRI with 1,018 samples), limiting their accuracies and reliability. In the past five years, several works have tailored for unsupervised representations of CT lesions via two-dimensional (2D) and three-dimensional (3D) self-supervised learning (SSL) algorithms. The 2D algorithms have difficulty capturing 3D information, and existing 3D algorithms are computationally heavy. Light-weight 3D SSL remains the boundary to explore. In this paper, we propose the spiral contrastive learning (SCL), which yields 3D representations in a computationally efficient manner. SCL first transforms 3D lesions to the 2D plane using an information-preserving spiral transformation, and then learn transformation-invariant features using 2D contrastive learning. For the augmentation, we consider natural image augmentations and medical image augmentations. We evaluate SCL by training a classification head upon the embedding layer. Experimental results show that SCL achieves state-of-the-art accuracy on LIDC-IDRI (89.72%), LNDb (82.09%) and TianChi (90.16%) for unsupervised representation learning. With 10% annotated data for fine-tune, the performance of SCL is comparable to that of supervised learning algorithms (85.75% vs. 85.03% on LIDC-IDRI, 78.20% vs. 73.44% on LNDb and 87.85% vs. 83.34% on TianChi, respectively). Meanwhile, SCL reduces the computational effort by 66.98% compared to other 3D SSL algorithms, demonstrating the efficiency of the proposed method in unsupervised pre-training." @default.
- W4320518829 created "2023-02-14" @default.
- W4320518829 creator A5006645897 @default.
- W4320518829 creator A5032712717 @default.
- W4320518829 creator A5047559861 @default.
- W4320518829 creator A5069142319 @default.
- W4320518829 creator A5089646011 @default.
- W4320518829 date "2022-08-22" @default.
- W4320518829 modified "2023-10-16" @default.
- W4320518829 title "Spiral Contrastive Learning: An Efficient 3D Representation Learning Method for Unannotated CT Lesions" @default.
- W4320518829 doi "https://doi.org/10.48550/arxiv.2208.10694" @default.
- W4320518829 hasPublicationYear "2022" @default.
- W4320518829 type Work @default.
- W4320518829 citedByCount "0" @default.
- W4320518829 crossrefType "posted-content" @default.
- W4320518829 hasAuthorship W4320518829A5006645897 @default.
- W4320518829 hasAuthorship W4320518829A5032712717 @default.
- W4320518829 hasAuthorship W4320518829A5047559861 @default.
- W4320518829 hasAuthorship W4320518829A5069142319 @default.
- W4320518829 hasAuthorship W4320518829A5089646011 @default.
- W4320518829 hasBestOaLocation W43205188291 @default.
- W4320518829 hasConcept C104317684 @default.
- W4320518829 hasConcept C11413529 @default.
- W4320518829 hasConcept C153180895 @default.
- W4320518829 hasConcept C154945302 @default.
- W4320518829 hasConcept C17744445 @default.
- W4320518829 hasConcept C185592680 @default.
- W4320518829 hasConcept C199539241 @default.
- W4320518829 hasConcept C204241405 @default.
- W4320518829 hasConcept C2776359362 @default.
- W4320518829 hasConcept C41008148 @default.
- W4320518829 hasConcept C41608201 @default.
- W4320518829 hasConcept C55493867 @default.
- W4320518829 hasConcept C59404180 @default.
- W4320518829 hasConcept C8038995 @default.
- W4320518829 hasConcept C94625758 @default.
- W4320518829 hasConceptScore W4320518829C104317684 @default.
- W4320518829 hasConceptScore W4320518829C11413529 @default.
- W4320518829 hasConceptScore W4320518829C153180895 @default.
- W4320518829 hasConceptScore W4320518829C154945302 @default.
- W4320518829 hasConceptScore W4320518829C17744445 @default.
- W4320518829 hasConceptScore W4320518829C185592680 @default.
- W4320518829 hasConceptScore W4320518829C199539241 @default.
- W4320518829 hasConceptScore W4320518829C204241405 @default.
- W4320518829 hasConceptScore W4320518829C2776359362 @default.
- W4320518829 hasConceptScore W4320518829C41008148 @default.
- W4320518829 hasConceptScore W4320518829C41608201 @default.
- W4320518829 hasConceptScore W4320518829C55493867 @default.
- W4320518829 hasConceptScore W4320518829C59404180 @default.
- W4320518829 hasConceptScore W4320518829C8038995 @default.
- W4320518829 hasConceptScore W4320518829C94625758 @default.
- W4320518829 hasLocation W43205188291 @default.
- W4320518829 hasOpenAccess W4320518829 @default.
- W4320518829 hasPrimaryLocation W43205188291 @default.
- W4320518829 hasRelatedWork W2592385986 @default.
- W4320518829 hasRelatedWork W2608244370 @default.
- W4320518829 hasRelatedWork W2695951553 @default.
- W4320518829 hasRelatedWork W2775464024 @default.
- W4320518829 hasRelatedWork W2899683012 @default.
- W4320518829 hasRelatedWork W2998168123 @default.
- W4320518829 hasRelatedWork W3111422391 @default.
- W4320518829 hasRelatedWork W3177029726 @default.
- W4320518829 hasRelatedWork W4287995534 @default.
- W4320518829 hasRelatedWork W4310745021 @default.
- W4320518829 isParatext "false" @default.
- W4320518829 isRetracted "false" @default.
- W4320518829 workType "article" @default.