Matches in SemOpenAlex for { <https://semopenalex.org/work/W3129330892> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W3129330892 endingPage "693" @default.
- W3129330892 startingPage "683" @default.
- W3129330892 abstract "Instance segmentation aims to locate targets in the image and segment each target at the pixel level, which is one of the most important tasks in computer vision. Mask R-CNN is a classic method of instance segmentation, but we find that its predicted masks are unclear and inaccurate near contours. To cope with this problem, we draw on the idea of contour matching based on distance transformation image and propose a novel loss function called contour loss. Contour loss is designed to specifically optimise the contour parts of the predicted masks, thus can assure more accurate instance segmentation. To make the proposed contour loss be jointly trained under modern neural network frameworks, we design a differentiable k-step distance transformation image calculation module, which can approximately compute truncated distance transformation images of the predicted mask and the corresponding ground-truth mask online. The proposed contour loss can be integrated into existing instance segmentation methods such as Mask R-CNN, and combined with their original loss functions without modification of the structures of inference network, thus has strong versatility. Experimental results on COCO show that contour loss is effective, which can further improve instance segmentation performances." @default.
- W3129330892 created "2021-03-01" @default.
- W3129330892 creator A5001500143 @default.
- W3129330892 creator A5029944675 @default.
- W3129330892 creator A5046839232 @default.
- W3129330892 creator A5084092583 @default.
- W3129330892 date "2022-06-06" @default.
- W3129330892 modified "2023-10-15" @default.
- W3129330892 title "Contour loss for instance segmentation via <i>k‐step</i> distance transformation image" @default.
- W3129330892 cites W1861492603 @default.
- W3129330892 cites W1903029394 @default.
- W3129330892 cites W2069537876 @default.
- W3129330892 cites W2071356085 @default.
- W3129330892 cites W2317851288 @default.
- W3129330892 cites W2791731612 @default.
- W3129330892 cites W2890862129 @default.
- W3129330892 cites W2963927307 @default.
- W3129330892 cites W3004308475 @default.
- W3129330892 cites W639708223 @default.
- W3129330892 doi "https://doi.org/10.1049/cvi2.12114" @default.
- W3129330892 hasPublicationYear "2022" @default.
- W3129330892 type Work @default.
- W3129330892 sameAs 3129330892 @default.
- W3129330892 citedByCount "0" @default.
- W3129330892 crossrefType "journal-article" @default.
- W3129330892 hasAuthorship W3129330892A5001500143 @default.
- W3129330892 hasAuthorship W3129330892A5029944675 @default.
- W3129330892 hasAuthorship W3129330892A5046839232 @default.
- W3129330892 hasAuthorship W3129330892A5084092583 @default.
- W3129330892 hasBestOaLocation W31293308922 @default.
- W3129330892 hasConcept C104317684 @default.
- W3129330892 hasConcept C115961682 @default.
- W3129330892 hasConcept C124504099 @default.
- W3129330892 hasConcept C14036430 @default.
- W3129330892 hasConcept C153180895 @default.
- W3129330892 hasConcept C154945302 @default.
- W3129330892 hasConcept C160633673 @default.
- W3129330892 hasConcept C185592680 @default.
- W3129330892 hasConcept C204241405 @default.
- W3129330892 hasConcept C31972630 @default.
- W3129330892 hasConcept C41008148 @default.
- W3129330892 hasConcept C55493867 @default.
- W3129330892 hasConcept C73621898 @default.
- W3129330892 hasConcept C78458016 @default.
- W3129330892 hasConcept C86803240 @default.
- W3129330892 hasConcept C89600930 @default.
- W3129330892 hasConceptScore W3129330892C104317684 @default.
- W3129330892 hasConceptScore W3129330892C115961682 @default.
- W3129330892 hasConceptScore W3129330892C124504099 @default.
- W3129330892 hasConceptScore W3129330892C14036430 @default.
- W3129330892 hasConceptScore W3129330892C153180895 @default.
- W3129330892 hasConceptScore W3129330892C154945302 @default.
- W3129330892 hasConceptScore W3129330892C160633673 @default.
- W3129330892 hasConceptScore W3129330892C185592680 @default.
- W3129330892 hasConceptScore W3129330892C204241405 @default.
- W3129330892 hasConceptScore W3129330892C31972630 @default.
- W3129330892 hasConceptScore W3129330892C41008148 @default.
- W3129330892 hasConceptScore W3129330892C55493867 @default.
- W3129330892 hasConceptScore W3129330892C73621898 @default.
- W3129330892 hasConceptScore W3129330892C78458016 @default.
- W3129330892 hasConceptScore W3129330892C86803240 @default.
- W3129330892 hasConceptScore W3129330892C89600930 @default.
- W3129330892 hasFunder F4320321001 @default.
- W3129330892 hasIssue "8" @default.
- W3129330892 hasLocation W31293308921 @default.
- W3129330892 hasLocation W31293308922 @default.
- W3129330892 hasOpenAccess W3129330892 @default.
- W3129330892 hasPrimaryLocation W31293308921 @default.
- W3129330892 hasRelatedWork W1507266234 @default.
- W3129330892 hasRelatedWork W1631910785 @default.
- W3129330892 hasRelatedWork W1669643531 @default.
- W3129330892 hasRelatedWork W1721780360 @default.
- W3129330892 hasRelatedWork W2110230079 @default.
- W3129330892 hasRelatedWork W2117664411 @default.
- W3129330892 hasRelatedWork W2117933325 @default.
- W3129330892 hasRelatedWork W2122581818 @default.
- W3129330892 hasRelatedWork W2159066190 @default.
- W3129330892 hasRelatedWork W2739874619 @default.
- W3129330892 hasVolume "16" @default.
- W3129330892 isParatext "false" @default.
- W3129330892 isRetracted "false" @default.
- W3129330892 magId "3129330892" @default.
- W3129330892 workType "article" @default.