FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2967759511> ?p ?o ?g. }

• W2967759511 abstract "Deep learning fostered a leap ahead in automated skin lesion analysis in the last two years. Those models are expensive to train and difficult to parameterize. Objective: We investigate methodological issues for designing and evaluating deep learning models for skin lesion analysis. We explore 10 choices faced by researchers: use of transfer learning, model architecture, train dataset, image resolution, type of data augmentation, input normalization, use of segmentation, duration of training, additional use of SVMs, and test data augmentation. Methods: We perform two full factorial experiments, for five different test datasets, resulting in 2560 exhaustive trials in our main experiment, and 1280 trials in our assessment of transfer learning. We analyze both with multi-way ANOVA. We use the exhaustive trials to simulate sequential decisions and ensembles, with and without the use of privileged information from the test set. Results -- main experiment: Amount of train data has disproportionate influence, explaining almost half the variation in performance. Of the other factors, test data augmentation and input resolution are the most influential. Deeper models, when combined, with extra data, also help. -- transfer experiment: Transfer learning is critical, its absence brings huge performance penalties. -- simulations: Ensembles of models are the best option to provide reliable results with limited resources, without using privileged information and sacrificing methodological rigor. Conclusions and Significance: Advancing research on automated skin lesion analysis requires curating larger public datasets. Indirect use of privileged information from the test set to design the models is a subtle, but frequent methodological mistake that leads to overoptimistic results. Ensembles of models are a cost-effective alternative to the expensive full-factorial and to the unstable sequential designs." @default.
• W2967759511 created "2019-08-22" @default.
• W2967759511 creator A5014063530 @default.
• W2967759511 creator A5045876000 @default.
• W2967759511 creator A5046548316 @default.
• W2967759511 creator A5051880838 @default.
• W2967759511 creator A5057680257 @default.
• W2967759511 creator A5063424641 @default.
• W2967759511 creator A5071026704 @default.
• W2967759511 date "2017-11-01" @default.
• W2967759511 modified "2023-10-17" @default.
• W2967759511 title "Data, Depth, and Design: Learning Reliable Models for Skin Lesion Analysis" @default.
• W2967759511 cites W1571052358 @default.
• W2967759511 cites W1901129140 @default.
• W2967759511 cites W1996828958 @default.
• W2967759511 cites W2031342017 @default.
• W2967759511 cites W2049277777 @default.
• W2967759511 cites W2061253660 @default.
• W2967759511 cites W2072546899 @default.
• W2967759511 cites W2099397584 @default.
• W2967759511 cites W2112865076 @default.
• W2967759511 cites W2180648740 @default.
• W2967759511 cites W2256917630 @default.
• W2967759511 cites W2274287116 @default.
• W2967759511 cites W2302255633 @default.
• W2967759511 cites W2338708481 @default.
• W2967759511 cites W2346705140 @default.
• W2967759511 cites W2426942631 @default.
• W2967759511 cites W2537189671 @default.
• W2967759511 cites W2564782580 @default.
• W2967759511 cites W2583697674 @default.
• W2967759511 cites W2591883108 @default.
• W2967759511 cites W2592236122 @default.
• W2967759511 cites W2592929672 @default.
• W2967759511 cites W2594652617 @default.
• W2967759511 cites W2594893939 @default.
• W2967759511 cites W2597394795 @default.
• W2967759511 cites W2602631755 @default.
• W2967759511 cites W2604384166 @default.
• W2967759511 cites W2612806369 @default.
• W2967759511 cites W2701556738 @default.
• W2967759511 cites W2753852778 @default.
• W2967759511 cites W2890655382 @default.
• W2967759511 cites W2963059730 @default.
• W2967759511 cites W2963946669 @default.
• W2967759511 cites W3102796228 @default.
• W2967759511 hasPublicationYear "2017" @default.
• W2967759511 type Work @default.
• W2967759511 sameAs 2967759511 @default.
• W2967759511 citedByCount "3" @default.
• W2967759511 countsByYear W29677595112018 @default.
• W2967759511 countsByYear W29677595112019 @default.
• W2967759511 crossrefType "posted-content" @default.
• W2967759511 hasAuthorship W2967759511A5014063530 @default.
• W2967759511 hasAuthorship W2967759511A5045876000 @default.
• W2967759511 hasAuthorship W2967759511A5046548316 @default.
• W2967759511 hasAuthorship W2967759511A5051880838 @default.
• W2967759511 hasAuthorship W2967759511A5057680257 @default.
• W2967759511 hasAuthorship W2967759511A5063424641 @default.
• W2967759511 hasAuthorship W2967759511A5071026704 @default.
• W2967759511 hasConcept C108583219 @default.
• W2967759511 hasConcept C119857082 @default.
• W2967759511 hasConcept C124101348 @default.
• W2967759511 hasConcept C136886441 @default.
• W2967759511 hasConcept C144024400 @default.
• W2967759511 hasConcept C150899416 @default.
• W2967759511 hasConcept C154945302 @default.
• W2967759511 hasConcept C16910744 @default.
• W2967759511 hasConcept C169903167 @default.
• W2967759511 hasConcept C177264268 @default.
• W2967759511 hasConcept C17744445 @default.
• W2967759511 hasConcept C19165224 @default.
• W2967759511 hasConcept C199360897 @default.
• W2967759511 hasConcept C199539241 @default.
• W2967759511 hasConcept C2777179996 @default.
• W2967759511 hasConcept C41008148 @default.
• W2967759511 hasConcept C58489278 @default.
• W2967759511 hasConcept C89600930 @default.
• W2967759511 hasConceptScore W2967759511C108583219 @default.
• W2967759511 hasConceptScore W2967759511C119857082 @default.
• W2967759511 hasConceptScore W2967759511C124101348 @default.
• W2967759511 hasConceptScore W2967759511C136886441 @default.
• W2967759511 hasConceptScore W2967759511C144024400 @default.
• W2967759511 hasConceptScore W2967759511C150899416 @default.
• W2967759511 hasConceptScore W2967759511C154945302 @default.
• W2967759511 hasConceptScore W2967759511C16910744 @default.
• W2967759511 hasConceptScore W2967759511C169903167 @default.
• W2967759511 hasConceptScore W2967759511C177264268 @default.
• W2967759511 hasConceptScore W2967759511C17744445 @default.
• W2967759511 hasConceptScore W2967759511C19165224 @default.
• W2967759511 hasConceptScore W2967759511C199360897 @default.
• W2967759511 hasConceptScore W2967759511C199539241 @default.
• W2967759511 hasConceptScore W2967759511C2777179996 @default.
• W2967759511 hasConceptScore W2967759511C41008148 @default.
• W2967759511 hasConceptScore W2967759511C58489278 @default.
• W2967759511 hasConceptScore W2967759511C89600930 @default.
• W2967759511 hasLocation W29677595111 @default.
• W2967759511 hasOpenAccess W2967759511 @default.
• W2967759511 hasPrimaryLocation W29677595111 @default.
• W2967759511 hasRelatedWork W2317236527 @default.

• next