FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4291237368> ?p ?o ?g. }

• W4291237368 endingPage "316" @default.
• W4291237368 startingPage "307" @default.
• W4291237368 abstract "BackgroundImage-based cancer classifiers suffer from a variety of problems which negatively affect their performance. For example, variation in image brightness or different cameras can already suffice to diminish performance. Ensemble solutions, where multiple model predictions are combined into one, can improve these problems. However, ensembles are computationally intensive and less transparent to practitioners than single model solutions. Constructing model soups, by averaging the weights of multiple models into a single model, could circumvent these limitations while still improving performance.ObjectiveTo investigate the performance of model soups for a dermoscopic melanoma-nevus skin cancer classification task with respect to (1) generalisation to images from other clinics, (2) robustness against small image changes and (3) calibration such that the confidences correspond closely to the actual predictive uncertainties.MethodsWe construct model soups by fine-tuning pre-trained models on seven different image resolutions and subsequently averaging their weights. Performance is evaluated on a multi-source dataset including holdout and external components.ResultsWe find that model soups improve generalisation and calibration on the external component while maintaining performance on the holdout component. For robustness, we observe performance improvements for pertubated test images, while the performance on corrupted test images remains on par.ConclusionsOverall, souping for skin cancer classifiers has a positive effect on generalisation, robustness and calibration. It is easy for practitioners to implement and by combining multiple models into a single model, complexity is reduced. This could be an important factor in achieving clinical applicability, as less complexity generally means more transparency." @default.
• W4291237368 created "2022-08-13" @default.
• W4291237368 creator A5007887862 @default.
• W4291237368 creator A5008568372 @default.
• W4291237368 creator A5008750058 @default.
• W4291237368 creator A5009718678 @default.
• W4291237368 creator A5017588184 @default.
• W4291237368 creator A5018377200 @default.
• W4291237368 creator A5025256131 @default.
• W4291237368 creator A5026100807 @default.
• W4291237368 creator A5030765130 @default.
• W4291237368 creator A5033111778 @default.
• W4291237368 creator A5034365278 @default.
• W4291237368 creator A5037516725 @default.
• W4291237368 creator A5040652983 @default.
• W4291237368 creator A5044736874 @default.
• W4291237368 creator A5046389977 @default.
• W4291237368 creator A5055619378 @default.
• W4291237368 creator A5056685282 @default.
• W4291237368 creator A5057226132 @default.
• W4291237368 creator A5058670875 @default.
• W4291237368 creator A5059931259 @default.
• W4291237368 creator A5063448654 @default.
• W4291237368 creator A5070424176 @default.
• W4291237368 creator A5072203765 @default.
• W4291237368 creator A5072592865 @default.
• W4291237368 creator A5073483894 @default.
• W4291237368 creator A5076078511 @default.
• W4291237368 creator A5076600698 @default.
• W4291237368 creator A5079922466 @default.
• W4291237368 creator A5084073390 @default.
• W4291237368 creator A5084788512 @default.
• W4291237368 creator A5086466437 @default.
• W4291237368 creator A5088740883 @default.
• W4291237368 creator A5091513635 @default.
• W4291237368 date "2022-09-01" @default.
• W4291237368 modified "2023-10-09" @default.
• W4291237368 title "Model soups improve performance of dermoscopic skin cancer classifiers" @default.
• W4291237368 cites W2581082771 @default.
• W4291237368 cites W2786147899 @default.
• W4291237368 cites W2797527544 @default.
• W4291237368 cites W2917303411 @default.
• W4291237368 cites W2921785317 @default.
• W4291237368 cites W2924551358 @default.
• W4291237368 cites W2937742783 @default.
• W4291237368 cites W2952971376 @default.
• W4291237368 cites W2968318837 @default.
• W4291237368 cites W2969096242 @default.
• W4291237368 cites W2972588473 @default.
• W4291237368 cites W2980349812 @default.
• W4291237368 cites W3014403957 @default.
• W4291237368 cites W3015351665 @default.
• W4291237368 cites W3036298167 @default.
• W4291237368 cites W3085870326 @default.
• W4291237368 cites W3102785203 @default.
• W4291237368 cites W3120507271 @default.
• W4291237368 cites W3121732873 @default.
• W4291237368 cites W3123777618 @default.
• W4291237368 cites W3158076558 @default.
• W4291237368 cites W3190348435 @default.
• W4291237368 cites W4225310474 @default.
• W4291237368 cites W4280517812 @default.
• W4291237368 cites W4289519099 @default.
• W4291237368 doi "https://doi.org/10.1016/j.ejca.2022.07.002" @default.
• W4291237368 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/35973360" @default.
• W4291237368 hasPublicationYear "2022" @default.
• W4291237368 type Work @default.
• W4291237368 citedByCount "1" @default.
• W4291237368 countsByYear W42912373682023 @default.
• W4291237368 crossrefType "journal-article" @default.
• W4291237368 hasAuthorship W4291237368A5007887862 @default.
• W4291237368 hasAuthorship W4291237368A5008568372 @default.
• W4291237368 hasAuthorship W4291237368A5008750058 @default.
• W4291237368 hasAuthorship W4291237368A5009718678 @default.
• W4291237368 hasAuthorship W4291237368A5017588184 @default.
• W4291237368 hasAuthorship W4291237368A5018377200 @default.
• W4291237368 hasAuthorship W4291237368A5025256131 @default.
• W4291237368 hasAuthorship W4291237368A5026100807 @default.
• W4291237368 hasAuthorship W4291237368A5030765130 @default.
• W4291237368 hasAuthorship W4291237368A5033111778 @default.
• W4291237368 hasAuthorship W4291237368A5034365278 @default.
• W4291237368 hasAuthorship W4291237368A5037516725 @default.
• W4291237368 hasAuthorship W4291237368A5040652983 @default.
• W4291237368 hasAuthorship W4291237368A5044736874 @default.
• W4291237368 hasAuthorship W4291237368A5046389977 @default.
• W4291237368 hasAuthorship W4291237368A5055619378 @default.
• W4291237368 hasAuthorship W4291237368A5056685282 @default.
• W4291237368 hasAuthorship W4291237368A5057226132 @default.
• W4291237368 hasAuthorship W4291237368A5058670875 @default.
• W4291237368 hasAuthorship W4291237368A5059931259 @default.
• W4291237368 hasAuthorship W4291237368A5063448654 @default.
• W4291237368 hasAuthorship W4291237368A5070424176 @default.
• W4291237368 hasAuthorship W4291237368A5072203765 @default.
• W4291237368 hasAuthorship W4291237368A5072592865 @default.
• W4291237368 hasAuthorship W4291237368A5073483894 @default.
• W4291237368 hasAuthorship W4291237368A5076078511 @default.
• W4291237368 hasAuthorship W4291237368A5076600698 @default.
• W4291237368 hasAuthorship W4291237368A5079922466 @default.

• next