Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313484562> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W4313484562 abstract "In this work, we focus on open vocabulary instance segmentation to expand a segmentation model to classify and segment instance-level novel categories. Previous approaches have relied on massive caption datasets and complex pipelines to establish one-to-one mappings between image regions and words in captions. However, such methods build noisy supervision by matching non-visible words to image regions, such as adjectives and verbs. Meanwhile, context words are also important for inferring the existence of novel objects as they show high inter-correlations with novel categories. To overcome these limitations, we devise a joint textbf{Caption Grounding and Generation (CGG)} framework, which incorporates a novel grounding loss that only focuses on matching object nouns to improve learning efficiency. We also introduce a caption generation head that enables additional supervision and contextual modeling as a complementation to the grounding loss. Our analysis and results demonstrate that grounding and generation components complement each other, significantly enhancing the segmentation performance for novel classes. Experiments on the COCO dataset with two settings: Open Vocabulary Instance Segmentation (OVIS) and Open Set Panoptic Segmentation (OSPS) demonstrate the superiority of the CGG. Specifically, CGG achieves a substantial improvement of 6.8% mAP for novel classes without extra data on the OVIS task and 15% PQ improvements for novel classes on the OSPS benchmark." @default.
- W4313484562 created "2023-01-06" @default.
- W4313484562 creator A5005626854 @default.
- W4313484562 creator A5024097240 @default.
- W4313484562 creator A5024765927 @default.
- W4313484562 creator A5036631624 @default.
- W4313484562 creator A5045854934 @default.
- W4313484562 creator A5049449805 @default.
- W4313484562 creator A5079738340 @default.
- W4313484562 date "2023-01-02" @default.
- W4313484562 modified "2023-09-27" @default.
- W4313484562 title "Betrayed by Captions: Joint Caption Grounding and Generation for Open Vocabulary Instance Segmentation" @default.
- W4313484562 doi "https://doi.org/10.48550/arxiv.2301.00805" @default.
- W4313484562 hasPublicationYear "2023" @default.
- W4313484562 type Work @default.
- W4313484562 citedByCount "0" @default.
- W4313484562 crossrefType "posted-content" @default.
- W4313484562 hasAuthorship W4313484562A5005626854 @default.
- W4313484562 hasAuthorship W4313484562A5024097240 @default.
- W4313484562 hasAuthorship W4313484562A5024765927 @default.
- W4313484562 hasAuthorship W4313484562A5036631624 @default.
- W4313484562 hasAuthorship W4313484562A5045854934 @default.
- W4313484562 hasAuthorship W4313484562A5049449805 @default.
- W4313484562 hasAuthorship W4313484562A5079738340 @default.
- W4313484562 hasBestOaLocation W43134845621 @default.
- W4313484562 hasConcept C105795698 @default.
- W4313484562 hasConcept C115961682 @default.
- W4313484562 hasConcept C120665830 @default.
- W4313484562 hasConcept C121332964 @default.
- W4313484562 hasConcept C13280743 @default.
- W4313484562 hasConcept C138885662 @default.
- W4313484562 hasConcept C151730666 @default.
- W4313484562 hasConcept C153180895 @default.
- W4313484562 hasConcept C154945302 @default.
- W4313484562 hasConcept C157657479 @default.
- W4313484562 hasConcept C162324750 @default.
- W4313484562 hasConcept C165064840 @default.
- W4313484562 hasConcept C170858558 @default.
- W4313484562 hasConcept C185798385 @default.
- W4313484562 hasConcept C187736073 @default.
- W4313484562 hasConcept C192209626 @default.
- W4313484562 hasConcept C204321447 @default.
- W4313484562 hasConcept C205649164 @default.
- W4313484562 hasConcept C2777601683 @default.
- W4313484562 hasConcept C2779343474 @default.
- W4313484562 hasConcept C2780451532 @default.
- W4313484562 hasConcept C33923547 @default.
- W4313484562 hasConcept C41008148 @default.
- W4313484562 hasConcept C41895202 @default.
- W4313484562 hasConcept C86803240 @default.
- W4313484562 hasConcept C89600930 @default.
- W4313484562 hasConceptScore W4313484562C105795698 @default.
- W4313484562 hasConceptScore W4313484562C115961682 @default.
- W4313484562 hasConceptScore W4313484562C120665830 @default.
- W4313484562 hasConceptScore W4313484562C121332964 @default.
- W4313484562 hasConceptScore W4313484562C13280743 @default.
- W4313484562 hasConceptScore W4313484562C138885662 @default.
- W4313484562 hasConceptScore W4313484562C151730666 @default.
- W4313484562 hasConceptScore W4313484562C153180895 @default.
- W4313484562 hasConceptScore W4313484562C154945302 @default.
- W4313484562 hasConceptScore W4313484562C157657479 @default.
- W4313484562 hasConceptScore W4313484562C162324750 @default.
- W4313484562 hasConceptScore W4313484562C165064840 @default.
- W4313484562 hasConceptScore W4313484562C170858558 @default.
- W4313484562 hasConceptScore W4313484562C185798385 @default.
- W4313484562 hasConceptScore W4313484562C187736073 @default.
- W4313484562 hasConceptScore W4313484562C192209626 @default.
- W4313484562 hasConceptScore W4313484562C204321447 @default.
- W4313484562 hasConceptScore W4313484562C205649164 @default.
- W4313484562 hasConceptScore W4313484562C2777601683 @default.
- W4313484562 hasConceptScore W4313484562C2779343474 @default.
- W4313484562 hasConceptScore W4313484562C2780451532 @default.
- W4313484562 hasConceptScore W4313484562C33923547 @default.
- W4313484562 hasConceptScore W4313484562C41008148 @default.
- W4313484562 hasConceptScore W4313484562C41895202 @default.
- W4313484562 hasConceptScore W4313484562C86803240 @default.
- W4313484562 hasConceptScore W4313484562C89600930 @default.
- W4313484562 hasLocation W43134845621 @default.
- W4313484562 hasLocation W43134845622 @default.
- W4313484562 hasOpenAccess W4313484562 @default.
- W4313484562 hasPrimaryLocation W43134845621 @default.
- W4313484562 hasRelatedWork W2025334826 @default.
- W4313484562 hasRelatedWork W2963877622 @default.
- W4313484562 hasRelatedWork W2971688299 @default.
- W4313484562 hasRelatedWork W3093267690 @default.
- W4313484562 hasRelatedWork W3119155061 @default.
- W4313484562 hasRelatedWork W3151908889 @default.
- W4313484562 hasRelatedWork W4287548803 @default.
- W4313484562 hasRelatedWork W4297080010 @default.
- W4313484562 hasRelatedWork W4306907108 @default.
- W4313484562 hasRelatedWork W4307074315 @default.
- W4313484562 isParatext "false" @default.
- W4313484562 isRetracted "false" @default.
- W4313484562 workType "article" @default.