Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386075582> ?p ?o ?g. }
- W4386075582 abstract "Feature matching is a challenging computer vision task that involves finding correspondences between two images of a 3D scene. In this paper we consider the dense approach instead of the more common sparse paradigm, thus striving to find all correspondences. Perhaps counter-intuitively, dense methods have previously shown inferior performance to their sparse and semi-sparse counterparts for estimation of two-view geometry. This changes with our novel dense method, which outperforms both dense and sparse methods on geometry estimation. The novelty is threefold: First, we propose a kernel regression global matcher. Secondly, we propose warp refinement through stacked feature maps and depthwise convolution kernels. Thirdly, we propose learning dense confidence through consistent depth and a balanced sampling approach for dense confidence maps. Through extensive experiments we confirm that our proposed dense method, Dense Kernelized Feature Matching, sets a new state-of-the-art on multiple geometry estimation benchmarks. In particular, we achieve an improvement on MegaDepth-1500 of +4.9 and +8.9 AUC@5° compared to the best previous sparse method and dense method respectively. Our code is provided at the following repository: https://github.com/Parskatt/DKM." @default.
- W4386075582 created "2023-08-23" @default.
- W4386075582 creator A5042087981 @default.
- W4386075582 creator A5052501521 @default.
- W4386075582 creator A5056741172 @default.
- W4386075582 creator A5070260526 @default.
- W4386075582 date "2023-06-01" @default.
- W4386075582 modified "2023-09-27" @default.
- W4386075582 title "DKM: Dense Kernelized Feature Matching for Geometry Estimation" @default.
- W4386075582 cites W1570870385 @default.
- W4386075582 cites W2025748541 @default.
- W4386075582 cites W2099927035 @default.
- W4386075582 cites W2117539524 @default.
- W4386075582 cites W2151103935 @default.
- W4386075582 cites W2194775991 @default.
- W4386075582 cites W2471962767 @default.
- W4386075582 cites W2594519801 @default.
- W4386075582 cites W2622594279 @default.
- W4386075582 cites W2740418457 @default.
- W4386075582 cites W2799166040 @default.
- W4386075582 cites W2897093986 @default.
- W4386075582 cites W2962705366 @default.
- W4386075582 cites W2963210849 @default.
- W4386075582 cites W2963446712 @default.
- W4386075582 cites W2963748588 @default.
- W4386075582 cites W2963760790 @default.
- W4386075582 cites W2979458572 @default.
- W4386075582 cites W3034275286 @default.
- W4386075582 cites W3035477606 @default.
- W4386075582 cites W3035564402 @default.
- W4386075582 cites W3041380886 @default.
- W4386075582 cites W3043075211 @default.
- W4386075582 cites W3094761124 @default.
- W4386075582 cites W3103648783 @default.
- W4386075582 cites W3106653913 @default.
- W4386075582 cites W3108164736 @default.
- W4386075582 cites W3166285241 @default.
- W4386075582 cites W3167674261 @default.
- W4386075582 cites W3173550280 @default.
- W4386075582 cites W3175295430 @default.
- W4386075582 cites W3194532238 @default.
- W4386075582 cites W3203518786 @default.
- W4386075582 cites W4200635831 @default.
- W4386075582 cites W4206212643 @default.
- W4386075582 cites W4211049957 @default.
- W4386075582 cites W4221154336 @default.
- W4386075582 cites W4292794046 @default.
- W4386075582 cites W4312247288 @default.
- W4386075582 cites W4312349930 @default.
- W4386075582 cites W4312583779 @default.
- W4386075582 cites W4312990699 @default.
- W4386075582 cites W4382458246 @default.
- W4386075582 doi "https://doi.org/10.1109/cvpr52729.2023.01704" @default.
- W4386075582 hasPublicationYear "2023" @default.
- W4386075582 type Work @default.
- W4386075582 citedByCount "1" @default.
- W4386075582 countsByYear W43860755822023 @default.
- W4386075582 crossrefType "proceedings-article" @default.
- W4386075582 hasAuthorship W4386075582A5042087981 @default.
- W4386075582 hasAuthorship W4386075582A5052501521 @default.
- W4386075582 hasAuthorship W4386075582A5056741172 @default.
- W4386075582 hasAuthorship W4386075582A5070260526 @default.
- W4386075582 hasConcept C105795698 @default.
- W4386075582 hasConcept C11413529 @default.
- W4386075582 hasConcept C114614502 @default.
- W4386075582 hasConcept C138885662 @default.
- W4386075582 hasConcept C153180895 @default.
- W4386075582 hasConcept C154945302 @default.
- W4386075582 hasConcept C165064840 @default.
- W4386075582 hasConcept C177264268 @default.
- W4386075582 hasConcept C199360897 @default.
- W4386075582 hasConcept C27206212 @default.
- W4386075582 hasConcept C2776401178 @default.
- W4386075582 hasConcept C2776760102 @default.
- W4386075582 hasConcept C2778738651 @default.
- W4386075582 hasConcept C33923547 @default.
- W4386075582 hasConcept C41008148 @default.
- W4386075582 hasConcept C41895202 @default.
- W4386075582 hasConcept C52622490 @default.
- W4386075582 hasConcept C74193536 @default.
- W4386075582 hasConceptScore W4386075582C105795698 @default.
- W4386075582 hasConceptScore W4386075582C11413529 @default.
- W4386075582 hasConceptScore W4386075582C114614502 @default.
- W4386075582 hasConceptScore W4386075582C138885662 @default.
- W4386075582 hasConceptScore W4386075582C153180895 @default.
- W4386075582 hasConceptScore W4386075582C154945302 @default.
- W4386075582 hasConceptScore W4386075582C165064840 @default.
- W4386075582 hasConceptScore W4386075582C177264268 @default.
- W4386075582 hasConceptScore W4386075582C199360897 @default.
- W4386075582 hasConceptScore W4386075582C27206212 @default.
- W4386075582 hasConceptScore W4386075582C2776401178 @default.
- W4386075582 hasConceptScore W4386075582C2776760102 @default.
- W4386075582 hasConceptScore W4386075582C2778738651 @default.
- W4386075582 hasConceptScore W4386075582C33923547 @default.
- W4386075582 hasConceptScore W4386075582C41008148 @default.
- W4386075582 hasConceptScore W4386075582C41895202 @default.
- W4386075582 hasConceptScore W4386075582C52622490 @default.
- W4386075582 hasConceptScore W4386075582C74193536 @default.
- W4386075582 hasFunder F4320322581 @default.
- W4386075582 hasLocation W43860755821 @default.