FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3098535516> ?p ?o ?g. }

• W3098535516 abstract "Most image segmentation algorithms are trained on binary masks formulated as a classification task per pixel. However, in applications such as medical imaging, this black-and-white approach is too constraining because the contrast between two tissues is often ill-defined, i.e., the voxels located on objects' edges contain a mixture of tissues. Consequently, assigning a single hard label can result in a detrimental approximation. Instead, a soft prediction containing non-binary values would overcome that limitation. We introduce SoftSeg, a deep learning training approach that takes advantage of soft ground truth labels, and is not bound to binary predictions. SoftSeg aims at solving a regression instead of a classification problem. This is achieved by using (i) no binarization after preprocessing and data augmentation, (ii) a normalized ReLU final activation layer (instead of sigmoid), and (iii) a regression loss function (instead of the traditional Dice loss). We assess the impact of these three features on three open-source MRI segmentation datasets from the spinal cord gray matter, the multiple sclerosis brain lesion, and the multimodal brain tumor segmentation challenges. Across multiple cross-validation iterations, SoftSeg outperformed the conventional approach, leading to an increase in Dice score of 2.0% on the gray matter dataset (p=0.001), 3.3% for the MS lesions, and 6.5% for the brain tumors. SoftSeg produces consistent soft predictions at tissues' interfaces and shows an increased sensitivity for small objects. The richness of soft labels could represent the inter-expert variability, the partial volume effect, and complement the model uncertainty estimation. The developed training pipeline can easily be incorporated into most of the existing deep learning architectures. It is already implemented in the freely-available deep learning toolbox ivadomed (this https URL)." @default.
• W3098535516 created "2020-11-23" @default.
• W3098535516 creator A5009765322 @default.
• W3098535516 creator A5026355368 @default.
• W3098535516 creator A5045086297 @default.
• W3098535516 date "2020-11-18" @default.
• W3098535516 modified "2023-10-15" @default.
• W3098535516 title "SoftSeg: Advantages of soft versus binary training for image segmentation" @default.
• W3098535516 cites W1588899353 @default.
• W3098535516 cites W1665214252 @default.
• W3098535516 cites W2014035569 @default.
• W3098535516 cites W2024050078 @default.
• W3098535516 cites W2095705004 @default.
• W3098535516 cites W2098979973 @default.
• W3098535516 cites W2117340355 @default.
• W3098535516 cites W2142592339 @default.
• W3098535516 cites W2148347694 @default.
• W3098535516 cites W2153480371 @default.
• W3098535516 cites W2183341477 @default.
• W3098535516 cites W2581082771 @default.
• W3098535516 cites W2592929672 @default.
• W3098535516 cites W2594770357 @default.
• W3098535516 cites W2621028221 @default.
• W3098535516 cites W2734349601 @default.
• W3098535516 cites W2755930428 @default.
• W3098535516 cites W2784195389 @default.
• W3098535516 cites W2792643794 @default.
• W3098535516 cites W2849179291 @default.
• W3098535516 cites W2937859967 @default.
• W3098535516 cites W2949117887 @default.
• W3098535516 cites W2950903920 @default.
• W3098535516 cites W2962914239 @default.
• W3098535516 cites W2963403868 @default.
• W3098535516 cites W2964059111 @default.
• W3098535516 cites W2966492512 @default.
• W3098535516 cites W2979410457 @default.
• W3098535516 cites W2981472752 @default.
• W3098535516 cites W2986198551 @default.
• W3098535516 cites W3030500365 @default.
• W3098535516 cites W3042259426 @default.
• W3098535516 cites W3091120243 @default.
• W3098535516 cites W3094554852 @default.
• W3098535516 cites W3098712157 @default.
• W3098535516 cites W3102961490 @default.
• W3098535516 hasPublicationYear "2020" @default.
• W3098535516 type Work @default.
• W3098535516 sameAs 3098535516 @default.
• W3098535516 citedByCount "1" @default.
• W3098535516 countsByYear W30985355162021 @default.
• W3098535516 crossrefType "posted-content" @default.
• W3098535516 hasAuthorship W3098535516A5009765322 @default.
• W3098535516 hasAuthorship W3098535516A5026355368 @default.
• W3098535516 hasAuthorship W3098535516A5045086297 @default.
• W3098535516 hasConcept C12267149 @default.
• W3098535516 hasConcept C124504099 @default.
• W3098535516 hasConcept C146849305 @default.
• W3098535516 hasConcept C153180895 @default.
• W3098535516 hasConcept C154945302 @default.
• W3098535516 hasConcept C160633673 @default.
• W3098535516 hasConcept C33923547 @default.
• W3098535516 hasConcept C34736171 @default.
• W3098535516 hasConcept C41008148 @default.
• W3098535516 hasConcept C48372109 @default.
• W3098535516 hasConcept C54170458 @default.
• W3098535516 hasConcept C66905080 @default.
• W3098535516 hasConcept C89600930 @default.
• W3098535516 hasConcept C94375191 @default.
• W3098535516 hasConceptScore W3098535516C12267149 @default.
• W3098535516 hasConceptScore W3098535516C124504099 @default.
• W3098535516 hasConceptScore W3098535516C146849305 @default.
• W3098535516 hasConceptScore W3098535516C153180895 @default.
• W3098535516 hasConceptScore W3098535516C154945302 @default.
• W3098535516 hasConceptScore W3098535516C160633673 @default.
• W3098535516 hasConceptScore W3098535516C33923547 @default.
• W3098535516 hasConceptScore W3098535516C34736171 @default.
• W3098535516 hasConceptScore W3098535516C41008148 @default.
• W3098535516 hasConceptScore W3098535516C48372109 @default.
• W3098535516 hasConceptScore W3098535516C54170458 @default.
• W3098535516 hasConceptScore W3098535516C66905080 @default.
• W3098535516 hasConceptScore W3098535516C89600930 @default.
• W3098535516 hasConceptScore W3098535516C94375191 @default.
• W3098535516 hasLocation W30985355161 @default.
• W3098535516 hasOpenAccess W3098535516 @default.
• W3098535516 hasPrimaryLocation W30985355161 @default.
• W3098535516 hasRelatedWork W2100199701 @default.
• W3098535516 hasRelatedWork W2146691973 @default.
• W3098535516 hasRelatedWork W2788335185 @default.
• W3098535516 hasRelatedWork W2788689866 @default.
• W3098535516 hasRelatedWork W2905091639 @default.
• W3098535516 hasRelatedWork W2913768941 @default.
• W3098535516 hasRelatedWork W2950671920 @default.
• W3098535516 hasRelatedWork W2979338721 @default.
• W3098535516 hasRelatedWork W2989082223 @default.
• W3098535516 hasRelatedWork W2993184403 @default.
• W3098535516 hasRelatedWork W2995779812 @default.
• W3098535516 hasRelatedWork W3006349040 @default.
• W3098535516 hasRelatedWork W3009445070 @default.
• W3098535516 hasRelatedWork W3014116974 @default.
• W3098535516 hasRelatedWork W3027226891 @default.
• W3098535516 hasRelatedWork W3092060218 @default.

• next