Matches in SemOpenAlex for { <https://semopenalex.org/work/W3136424010> ?p ?o ?g. }
- W3136424010 endingPage "102035" @default.
- W3136424010 startingPage "102035" @default.
- W3136424010 abstract "The loss function is an important component in deep learning-based segmentation methods. Over the past five years, many loss functions have been proposed for various segmentation tasks. However, a systematic study of the utility of these loss functions is missing. In this paper, we present a comprehensive review of segmentation loss functions in an organized manner. We also conduct the first large-scale analysis of 20 general loss functions on four typical 3D segmentation tasks involving six public datasets from 10+ medical centers. The results show that none of the losses can consistently achieve the best performance on the four segmentation tasks, but compound loss functions (e.g. Dice with TopK loss, focal loss, Hausdorff distance loss, and boundary loss) are the most robust losses. Our code and segmentation results are publicly available and can serve as a loss function benchmark. We hope this work will also provide insights on new loss function development for the community." @default.
- W3136424010 created "2021-03-29" @default.
- W3136424010 creator A5005173070 @default.
- W3136424010 creator A5022571340 @default.
- W3136424010 creator A5029260213 @default.
- W3136424010 creator A5032447166 @default.
- W3136424010 creator A5038819595 @default.
- W3136424010 creator A5049872422 @default.
- W3136424010 creator A5052832070 @default.
- W3136424010 creator A5054988731 @default.
- W3136424010 date "2021-07-01" @default.
- W3136424010 modified "2023-10-16" @default.
- W3136424010 title "Loss odyssey in medical image segmentation" @default.
- W3136424010 cites W1641498739 @default.
- W3136424010 cites W2791680898 @default.
- W3136424010 cites W2804047627 @default.
- W3136424010 cites W2888358068 @default.
- W3136424010 cites W2900237898 @default.
- W3136424010 cites W2903617078 @default.
- W3136424010 cites W2928133111 @default.
- W3136424010 cites W2962731543 @default.
- W3136424010 cites W2963043051 @default.
- W3136424010 cites W2964227007 @default.
- W3136424010 cites W2966333895 @default.
- W3136424010 cites W2966434031 @default.
- W3136424010 cites W2970367578 @default.
- W3136424010 cites W3035546112 @default.
- W3136424010 cites W3090605478 @default.
- W3136424010 cites W3093458154 @default.
- W3136424010 cites W3112701542 @default.
- W3136424010 doi "https://doi.org/10.1016/j.media.2021.102035" @default.
- W3136424010 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/33813286" @default.
- W3136424010 hasPublicationYear "2021" @default.
- W3136424010 type Work @default.
- W3136424010 sameAs 3136424010 @default.
- W3136424010 citedByCount "176" @default.
- W3136424010 countsByYear W31364240102012 @default.
- W3136424010 countsByYear W31364240102020 @default.
- W3136424010 countsByYear W31364240102021 @default.
- W3136424010 countsByYear W31364240102022 @default.
- W3136424010 countsByYear W31364240102023 @default.
- W3136424010 crossrefType "journal-article" @default.
- W3136424010 hasAuthorship W3136424010A5005173070 @default.
- W3136424010 hasAuthorship W3136424010A5022571340 @default.
- W3136424010 hasAuthorship W3136424010A5029260213 @default.
- W3136424010 hasAuthorship W3136424010A5032447166 @default.
- W3136424010 hasAuthorship W3136424010A5038819595 @default.
- W3136424010 hasAuthorship W3136424010A5049872422 @default.
- W3136424010 hasAuthorship W3136424010A5052832070 @default.
- W3136424010 hasAuthorship W3136424010A5054988731 @default.
- W3136424010 hasConcept C104317684 @default.
- W3136424010 hasConcept C105795698 @default.
- W3136424010 hasConcept C105951970 @default.
- W3136424010 hasConcept C124504099 @default.
- W3136424010 hasConcept C127716648 @default.
- W3136424010 hasConcept C134306372 @default.
- W3136424010 hasConcept C14036430 @default.
- W3136424010 hasConcept C141898687 @default.
- W3136424010 hasConcept C154945302 @default.
- W3136424010 hasConcept C185592680 @default.
- W3136424010 hasConcept C185798385 @default.
- W3136424010 hasConcept C205649164 @default.
- W3136424010 hasConcept C22029948 @default.
- W3136424010 hasConcept C33923547 @default.
- W3136424010 hasConcept C41008148 @default.
- W3136424010 hasConcept C55493867 @default.
- W3136424010 hasConcept C58640448 @default.
- W3136424010 hasConcept C62354387 @default.
- W3136424010 hasConcept C78458016 @default.
- W3136424010 hasConcept C86803240 @default.
- W3136424010 hasConcept C89600930 @default.
- W3136424010 hasConceptScore W3136424010C104317684 @default.
- W3136424010 hasConceptScore W3136424010C105795698 @default.
- W3136424010 hasConceptScore W3136424010C105951970 @default.
- W3136424010 hasConceptScore W3136424010C124504099 @default.
- W3136424010 hasConceptScore W3136424010C127716648 @default.
- W3136424010 hasConceptScore W3136424010C134306372 @default.
- W3136424010 hasConceptScore W3136424010C14036430 @default.
- W3136424010 hasConceptScore W3136424010C141898687 @default.
- W3136424010 hasConceptScore W3136424010C154945302 @default.
- W3136424010 hasConceptScore W3136424010C185592680 @default.
- W3136424010 hasConceptScore W3136424010C185798385 @default.
- W3136424010 hasConceptScore W3136424010C205649164 @default.
- W3136424010 hasConceptScore W3136424010C22029948 @default.
- W3136424010 hasConceptScore W3136424010C33923547 @default.
- W3136424010 hasConceptScore W3136424010C41008148 @default.
- W3136424010 hasConceptScore W3136424010C55493867 @default.
- W3136424010 hasConceptScore W3136424010C58640448 @default.
- W3136424010 hasConceptScore W3136424010C62354387 @default.
- W3136424010 hasConceptScore W3136424010C78458016 @default.
- W3136424010 hasConceptScore W3136424010C86803240 @default.
- W3136424010 hasConceptScore W3136424010C89600930 @default.
- W3136424010 hasFunder F4320321001 @default.
- W3136424010 hasLocation W31364240101 @default.
- W3136424010 hasLocation W31364240102 @default.
- W3136424010 hasOpenAccess W3136424010 @default.
- W3136424010 hasPrimaryLocation W31364240101 @default.
- W3136424010 hasRelatedWork W2138214894 @default.