Matches in SemOpenAlex for { <https://semopenalex.org/work/W4320041089> ?p ?o ?g. }
Showing items 1 to 65 of
65
with 100 items per page.
- W4320041089 endingPage "64" @default.
- W4320041089 startingPage "52" @default.
- W4320041089 abstract "For the Diabetic Foot Ulcer Challenge 2022 (DFUC2022) hosted by MICCAI 2022, we built a machine learning model based on the architecture of TransFuse [20] to accomplish the segmentation task. The TransFuse model combines Transformers and convolutional neural networks (CNNs), taking advantage of both local and global features. In this paper, we propose a modification to the data flow in encoder necks for decoding features in the higher resolution level, and in fusion modules for more efficient attention. Furthermore, to minimize the information loss as a result of resizing, we propose new techniques in both training and testing algorithms. Firstly, a region proposal network (RPN) is introduced from object detection methods and is used at the image pre-processing phase. It crops fixed size images from origin images, so that the high resolution input can be fed into TransFuse. We also applied test-time augmentation following a similar concept to RPN. We crop fixed size images at each corner and use edge pooling to ensemble them properly." @default.
- W4320041089 created "2023-02-12" @default.
- W4320041089 creator A5065247900 @default.
- W4320041089 creator A5084640645 @default.
- W4320041089 creator A5091713952 @default.
- W4320041089 date "2023-01-01" @default.
- W4320041089 modified "2023-10-14" @default.
- W4320041089 title "Capture the Devil in the Details via Partition-then-Ensemble on Higher Resolution Images" @default.
- W4320041089 cites W1536680647 @default.
- W4320041089 cites W2194775991 @default.
- W4320041089 cites W2752782242 @default.
- W4320041089 cites W2884585870 @default.
- W4320041089 cites W2890781352 @default.
- W4320041089 cites W2963351448 @default.
- W4320041089 cites W2964274014 @default.
- W4320041089 cites W2967307920 @default.
- W4320041089 cites W2969913432 @default.
- W4320041089 cites W2987175876 @default.
- W4320041089 cites W3092344722 @default.
- W4320041089 cites W3138516171 @default.
- W4320041089 cites W3188036383 @default.
- W4320041089 cites W3204166336 @default.
- W4320041089 cites W4313007769 @default.
- W4320041089 doi "https://doi.org/10.1007/978-3-031-26354-5_5" @default.
- W4320041089 hasPublicationYear "2023" @default.
- W4320041089 type Work @default.
- W4320041089 citedByCount "0" @default.
- W4320041089 crossrefType "book-chapter" @default.
- W4320041089 hasAuthorship W4320041089A5065247900 @default.
- W4320041089 hasAuthorship W4320041089A5084640645 @default.
- W4320041089 hasAuthorship W4320041089A5091713952 @default.
- W4320041089 hasConcept C111919701 @default.
- W4320041089 hasConcept C118505674 @default.
- W4320041089 hasConcept C153180895 @default.
- W4320041089 hasConcept C154945302 @default.
- W4320041089 hasConcept C41008148 @default.
- W4320041089 hasConcept C70437156 @default.
- W4320041089 hasConcept C81363708 @default.
- W4320041089 hasConcept C89600930 @default.
- W4320041089 hasConceptScore W4320041089C111919701 @default.
- W4320041089 hasConceptScore W4320041089C118505674 @default.
- W4320041089 hasConceptScore W4320041089C153180895 @default.
- W4320041089 hasConceptScore W4320041089C154945302 @default.
- W4320041089 hasConceptScore W4320041089C41008148 @default.
- W4320041089 hasConceptScore W4320041089C70437156 @default.
- W4320041089 hasConceptScore W4320041089C81363708 @default.
- W4320041089 hasConceptScore W4320041089C89600930 @default.
- W4320041089 hasLocation W43200410891 @default.
- W4320041089 hasOpenAccess W4320041089 @default.
- W4320041089 hasPrimaryLocation W43200410891 @default.
- W4320041089 hasRelatedWork W2291847203 @default.
- W4320041089 hasRelatedWork W2424871898 @default.
- W4320041089 hasRelatedWork W2514274290 @default.
- W4320041089 hasRelatedWork W2517027266 @default.
- W4320041089 hasRelatedWork W2758063741 @default.
- W4320041089 hasRelatedWork W2940661641 @default.
- W4320041089 hasRelatedWork W2951391129 @default.
- W4320041089 hasRelatedWork W2969680539 @default.
- W4320041089 hasRelatedWork W3172946746 @default.
- W4320041089 hasRelatedWork W4308191152 @default.
- W4320041089 isParatext "false" @default.
- W4320041089 isRetracted "false" @default.
- W4320041089 workType "book-chapter" @default.