Matches in SemOpenAlex for { <https://semopenalex.org/work/W3096850306> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W3096850306 abstract "Segmentation of Magnetic Resonance Images (MRI) of abdominal organs is useful for analysis prior to surgical procedures and for further processing. Deep Learning (DL) has become the standard, researchers have proposed improvements that include multiple views, ensembles and voting. Loss function alternatives, while being crucial to guide automated learning, have not been compared in detail. In this work we analyze limitations of popular metrics and their use as loss, study alternative loss variations based on those and other modifications and search for the best approach. An experimental setup was necessary to assess the alternatives. Results for the top scoring network and top scoring loss show improvements between 2 and 11 percentage points (pp) in Jaccard Index (JI), depending on organ and patient (sequence), for a total of 22 pp over 4 organs, all this being obtained just by choosing the best performing loss function instead of cross-entropy or dice. Our results apply directly to MRI of abdominal organs, with important practical implications for other architectures, as they can be applied easily to any of them. They also show the worth of variants of loss function and loss tuning, with future work needed to generalize and test in other contexts." @default.
- W3096850306 created "2020-11-09" @default.
- W3096850306 creator A5049382757 @default.
- W3096850306 date "2020-11-02" @default.
- W3096850306 modified "2023-09-25" @default.
- W3096850306 title "Investigation of Loss Functions for Improving Deep Segmentation of Abdominal Organs from MRI" @default.
- W3096850306 doi "https://doi.org/10.20944/preprints202011.0023.v1" @default.
- W3096850306 hasPublicationYear "2020" @default.
- W3096850306 type Work @default.
- W3096850306 sameAs 3096850306 @default.
- W3096850306 citedByCount "0" @default.
- W3096850306 crossrefType "posted-content" @default.
- W3096850306 hasAuthorship W3096850306A5049382757 @default.
- W3096850306 hasBestOaLocation W30968503061 @default.
- W3096850306 hasConcept C104317684 @default.
- W3096850306 hasConcept C105795698 @default.
- W3096850306 hasConcept C105951970 @default.
- W3096850306 hasConcept C108583219 @default.
- W3096850306 hasConcept C119857082 @default.
- W3096850306 hasConcept C127716648 @default.
- W3096850306 hasConcept C14036430 @default.
- W3096850306 hasConcept C153180895 @default.
- W3096850306 hasConcept C154945302 @default.
- W3096850306 hasConcept C167981619 @default.
- W3096850306 hasConcept C17744445 @default.
- W3096850306 hasConcept C185592680 @default.
- W3096850306 hasConcept C199539241 @default.
- W3096850306 hasConcept C203519979 @default.
- W3096850306 hasConcept C22029948 @default.
- W3096850306 hasConcept C33923547 @default.
- W3096850306 hasConcept C41008148 @default.
- W3096850306 hasConcept C520049643 @default.
- W3096850306 hasConcept C55493867 @default.
- W3096850306 hasConcept C78458016 @default.
- W3096850306 hasConcept C86803240 @default.
- W3096850306 hasConcept C89600930 @default.
- W3096850306 hasConcept C94625758 @default.
- W3096850306 hasConceptScore W3096850306C104317684 @default.
- W3096850306 hasConceptScore W3096850306C105795698 @default.
- W3096850306 hasConceptScore W3096850306C105951970 @default.
- W3096850306 hasConceptScore W3096850306C108583219 @default.
- W3096850306 hasConceptScore W3096850306C119857082 @default.
- W3096850306 hasConceptScore W3096850306C127716648 @default.
- W3096850306 hasConceptScore W3096850306C14036430 @default.
- W3096850306 hasConceptScore W3096850306C153180895 @default.
- W3096850306 hasConceptScore W3096850306C154945302 @default.
- W3096850306 hasConceptScore W3096850306C167981619 @default.
- W3096850306 hasConceptScore W3096850306C17744445 @default.
- W3096850306 hasConceptScore W3096850306C185592680 @default.
- W3096850306 hasConceptScore W3096850306C199539241 @default.
- W3096850306 hasConceptScore W3096850306C203519979 @default.
- W3096850306 hasConceptScore W3096850306C22029948 @default.
- W3096850306 hasConceptScore W3096850306C33923547 @default.
- W3096850306 hasConceptScore W3096850306C41008148 @default.
- W3096850306 hasConceptScore W3096850306C520049643 @default.
- W3096850306 hasConceptScore W3096850306C55493867 @default.
- W3096850306 hasConceptScore W3096850306C78458016 @default.
- W3096850306 hasConceptScore W3096850306C86803240 @default.
- W3096850306 hasConceptScore W3096850306C89600930 @default.
- W3096850306 hasConceptScore W3096850306C94625758 @default.
- W3096850306 hasLocation W30968503061 @default.
- W3096850306 hasOpenAccess W3096850306 @default.
- W3096850306 hasPrimaryLocation W30968503061 @default.
- W3096850306 hasRelatedWork W3129629 @default.
- W3096850306 hasRelatedWork W4500673 @default.
- W3096850306 hasRelatedWork W5046839 @default.
- W3096850306 hasRelatedWork W5053613 @default.
- W3096850306 hasRelatedWork W5275551 @default.
- W3096850306 hasRelatedWork W553262 @default.
- W3096850306 hasRelatedWork W692962 @default.
- W3096850306 hasRelatedWork W8695 @default.
- W3096850306 hasRelatedWork W8815345 @default.
- W3096850306 hasRelatedWork W8848861 @default.
- W3096850306 isParatext "false" @default.
- W3096850306 isRetracted "false" @default.
- W3096850306 magId "3096850306" @default.
- W3096850306 workType "article" @default.