Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313590873> ?p ?o ?g. }
- W4313590873 endingPage "15" @default.
- W4313590873 startingPage "1" @default.
- W4313590873 abstract "This paper proposes a novel paradigm for the unsupervised learning of object landmark detectors. Contrary to existing methods that build on auxiliary tasks such as image generation or equivariance, we propose a self-training approach where, departing from generic keypoints, a landmark detector and descriptor is trained to improve itself, tuning the keypoints into distinctive landmarks. To this end, we propose an iterative algorithm that alternates between producing new pseudo-labels through feature clustering and learning distinctive features for each pseudo-class through contrastive learning. With a shared backbone for the landmark detector and descriptor, the keypoint locations progressively converge to stable landmarks, filtering those less stable. Compared to previous works, our approach can learn points that are more flexible in terms of capturing large viewpoint changes. We validate our method on a variety of difficult datasets, including LS3D, BBCPose, Human3.6M and PennAction, achieving new state of the art results. Code and models can be found at https://github.com/dimitrismallis/KeypointsToLandmarks/." @default.
- W4313590873 created "2023-01-06" @default.
- W4313590873 creator A5021857035 @default.
- W4313590873 creator A5023727222 @default.
- W4313590873 creator A5024224610 @default.
- W4313590873 creator A5085591448 @default.
- W4313590873 date "2023-01-01" @default.
- W4313590873 modified "2023-09-26" @default.
- W4313590873 title "From Keypoints to Object Landmarks via Self-Training Correspondence: A novel approach to Unsupervised Landmark Discovery" @default.
- W4313590873 cites W1495267108 @default.
- W4313590873 cites W1677409904 @default.
- W4313590873 cites W1834627138 @default.
- W4313590873 cites W1896424170 @default.
- W4313590873 cites W1994529670 @default.
- W4313590873 cites W2011700682 @default.
- W4313590873 cites W2012885984 @default.
- W4313590873 cites W2058961190 @default.
- W4313590873 cites W2101032778 @default.
- W4313590873 cites W2103972604 @default.
- W4313590873 cites W2117228865 @default.
- W4313590873 cites W2124386111 @default.
- W4313590873 cites W2136703000 @default.
- W4313590873 cites W2137591992 @default.
- W4313590873 cites W2147334734 @default.
- W4313590873 cites W2151103935 @default.
- W4313590873 cites W2166344663 @default.
- W4313590873 cites W2222512263 @default.
- W4313590873 cites W2251810906 @default.
- W4313590873 cites W2307770531 @default.
- W4313590873 cites W2320444803 @default.
- W4313590873 cites W2321533354 @default.
- W4313590873 cites W2322739735 @default.
- W4313590873 cites W2329995605 @default.
- W4313590873 cites W2345643369 @default.
- W4313590873 cites W2518754566 @default.
- W4313590873 cites W2523270796 @default.
- W4313590873 cites W2552414813 @default.
- W4313590873 cites W2583837639 @default.
- W4313590873 cites W2583894904 @default.
- W4313590873 cites W2620629206 @default.
- W4313590873 cites W2777511827 @default.
- W4313590873 cites W2793720103 @default.
- W4313590873 cites W2807725536 @default.
- W4313590873 cites W2883725317 @default.
- W4313590873 cites W2913429812 @default.
- W4313590873 cites W2962925415 @default.
- W4313590873 cites W2962981304 @default.
- W4313590873 cites W2963168844 @default.
- W4313590873 cites W2963402313 @default.
- W4313590873 cites W2963419579 @default.
- W4313590873 cites W2963465221 @default.
- W4313590873 cites W2963583792 @default.
- W4313590873 cites W2964014798 @default.
- W4313590873 cites W2964352379 @default.
- W4313590873 cites W2982697283 @default.
- W4313590873 cites W2987936369 @default.
- W4313590873 cites W2998388430 @default.
- W4313590873 cites W3000817459 @default.
- W4313590873 cites W3034345981 @default.
- W4313590873 cites W3035160371 @default.
- W4313590873 cites W3043075211 @default.
- W4313590873 cites W3102288316 @default.
- W4313590873 cites W3104792420 @default.
- W4313590873 cites W602397586 @default.
- W4313590873 cites W639708223 @default.
- W4313590873 doi "https://doi.org/10.1109/tpami.2023.3234212" @default.
- W4313590873 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37018262" @default.
- W4313590873 hasPublicationYear "2023" @default.
- W4313590873 type Work @default.
- W4313590873 citedByCount "1" @default.
- W4313590873 countsByYear W43135908732023 @default.
- W4313590873 crossrefType "journal-article" @default.
- W4313590873 hasAuthorship W4313590873A5021857035 @default.
- W4313590873 hasAuthorship W4313590873A5023727222 @default.
- W4313590873 hasAuthorship W4313590873A5024224610 @default.
- W4313590873 hasAuthorship W4313590873A5085591448 @default.
- W4313590873 hasBestOaLocation W43135908732 @default.
- W4313590873 hasConcept C115961682 @default.
- W4313590873 hasConcept C138885662 @default.
- W4313590873 hasConcept C153180895 @default.
- W4313590873 hasConcept C154945302 @default.
- W4313590873 hasConcept C177264268 @default.
- W4313590873 hasConcept C199360897 @default.
- W4313590873 hasConcept C2776401178 @default.
- W4313590873 hasConcept C2776760102 @default.
- W4313590873 hasConcept C2780297707 @default.
- W4313590873 hasConcept C2781238097 @default.
- W4313590873 hasConcept C31972630 @default.
- W4313590873 hasConcept C41008148 @default.
- W4313590873 hasConcept C41895202 @default.
- W4313590873 hasConcept C73555534 @default.
- W4313590873 hasConcept C76155785 @default.
- W4313590873 hasConcept C8038995 @default.
- W4313590873 hasConcept C94915269 @default.
- W4313590873 hasConceptScore W4313590873C115961682 @default.
- W4313590873 hasConceptScore W4313590873C138885662 @default.
- W4313590873 hasConceptScore W4313590873C153180895 @default.
- W4313590873 hasConceptScore W4313590873C154945302 @default.