Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386075493> ?p ?o ?g. }
- W4386075493 abstract "In this work, instead of directly predicting the pixel-level segmentation masks, the problem of referring image seg-mentation is formulated as sequential polygon generation, and the predicted polygons can be later converted into segmentation masks. This is enabled by a new sequence-to-sequence framework, Polygon Transformer (PolyFormer), which takes a sequence of image patches and text query to-kens as input, and outputs a sequence of polygon vertices autoregressively. For more accurate geometric localization, we propose a regression-based decoder, which predicts the precise floating-point coordinates directly, without any co-ordinate quantization error. In the experiments, PolyFormer outperforms the prior art by a clear margin, e.g., 5.40% and 4.52% absolute improvements on the challenging Re-fCOCO+ and RefCOCOg datasets. It also shows strong generalization ability when evaluated on the referring video segmentation task without fine-tuning, e.g., achieving competitive 61.5% J&F on the Ref-DAVIS17 dataset." @default.
- W4386075493 created "2023-08-23" @default.
- W4386075493 creator A5004658590 @default.
- W4386075493 creator A5024623500 @default.
- W4386075493 creator A5034282541 @default.
- W4386075493 creator A5037560489 @default.
- W4386075493 creator A5053816690 @default.
- W4386075493 creator A5059422823 @default.
- W4386075493 creator A5084058732 @default.
- W4386075493 date "2023-06-01" @default.
- W4386075493 modified "2023-10-16" @default.
- W4386075493 title "PolyFormer: Referring Image Segmentation as Sequential Polygon Generation" @default.
- W4386075493 cites W1536680647 @default.
- W4386075493 cites W2006147162 @default.
- W4386075493 cites W2064675550 @default.
- W4386075493 cites W2110764733 @default.
- W4386075493 cites W2194775991 @default.
- W4386075493 cites W2216125271 @default.
- W4386075493 cites W2251512949 @default.
- W4386075493 cites W2277195237 @default.
- W4386075493 cites W2558535589 @default.
- W4386075493 cites W2605127024 @default.
- W4386075493 cites W2609825896 @default.
- W4386075493 cites W2770129969 @default.
- W4386075493 cites W2787287818 @default.
- W4386075493 cites W2798556392 @default.
- W4386075493 cites W2946086442 @default.
- W4386075493 cites W2962764817 @default.
- W4386075493 cites W2963109634 @default.
- W4386075493 cites W2963150697 @default.
- W4386075493 cites W2963296245 @default.
- W4386075493 cites W2963857746 @default.
- W4386075493 cites W2964199361 @default.
- W4386075493 cites W2964241181 @default.
- W4386075493 cites W2964345792 @default.
- W4386075493 cites W2970231061 @default.
- W4386075493 cites W2987734933 @default.
- W4386075493 cites W2993182889 @default.
- W4386075493 cites W3004019157 @default.
- W4386075493 cites W3034325957 @default.
- W4386075493 cites W3034512672 @default.
- W4386075493 cites W3034692043 @default.
- W4386075493 cites W3034727271 @default.
- W4386075493 cites W3034772468 @default.
- W4386075493 cites W3034826836 @default.
- W4386075493 cites W3035097537 @default.
- W4386075493 cites W3035709993 @default.
- W4386075493 cites W3093025045 @default.
- W4386075493 cites W3104844437 @default.
- W4386075493 cites W3138516171 @default.
- W4386075493 cites W3159619744 @default.
- W4386075493 cites W3169998662 @default.
- W4386075493 cites W3172522282 @default.
- W4386075493 cites W3173220247 @default.
- W4386075493 cites W3187664142 @default.
- W4386075493 cites W3201770677 @default.
- W4386075493 cites W3216551675 @default.
- W4386075493 cites W4200631575 @default.
- W4386075493 cites W4226024706 @default.
- W4386075493 cites W4312543911 @default.
- W4386075493 cites W4312691946 @default.
- W4386075493 cites W4312933868 @default.
- W4386075493 doi "https://doi.org/10.1109/cvpr52729.2023.01789" @default.
- W4386075493 hasPublicationYear "2023" @default.
- W4386075493 type Work @default.
- W4386075493 citedByCount "1" @default.
- W4386075493 countsByYear W43860754932023 @default.
- W4386075493 crossrefType "proceedings-article" @default.
- W4386075493 hasAuthorship W4386075493A5004658590 @default.
- W4386075493 hasAuthorship W4386075493A5024623500 @default.
- W4386075493 hasAuthorship W4386075493A5034282541 @default.
- W4386075493 hasAuthorship W4386075493A5037560489 @default.
- W4386075493 hasAuthorship W4386075493A5053816690 @default.
- W4386075493 hasAuthorship W4386075493A5059422823 @default.
- W4386075493 hasAuthorship W4386075493A5084058732 @default.
- W4386075493 hasConcept C11413529 @default.
- W4386075493 hasConcept C121684516 @default.
- W4386075493 hasConcept C124504099 @default.
- W4386075493 hasConcept C126042441 @default.
- W4386075493 hasConcept C153180895 @default.
- W4386075493 hasConcept C154945302 @default.
- W4386075493 hasConcept C160633673 @default.
- W4386075493 hasConcept C190694206 @default.
- W4386075493 hasConcept C2778112365 @default.
- W4386075493 hasConcept C28855332 @default.
- W4386075493 hasConcept C31487907 @default.
- W4386075493 hasConcept C31972630 @default.
- W4386075493 hasConcept C41008148 @default.
- W4386075493 hasConcept C54355233 @default.
- W4386075493 hasConcept C59902600 @default.
- W4386075493 hasConcept C65885262 @default.
- W4386075493 hasConcept C76155785 @default.
- W4386075493 hasConcept C86803240 @default.
- W4386075493 hasConcept C89600930 @default.
- W4386075493 hasConceptScore W4386075493C11413529 @default.
- W4386075493 hasConceptScore W4386075493C121684516 @default.
- W4386075493 hasConceptScore W4386075493C124504099 @default.
- W4386075493 hasConceptScore W4386075493C126042441 @default.
- W4386075493 hasConceptScore W4386075493C153180895 @default.
- W4386075493 hasConceptScore W4386075493C154945302 @default.