Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313003742> ?p ?o ?g. }
- W4313003742 endingPage "230" @default.
- W4313003742 startingPage "212" @default.
- W4313003742 abstract "AbstractFully supervised human mesh recovery methods are data-hungry and have poor generalizability due to the limited availability and diversity of 3D-annotated benchmark datasets. Recent progress in self-supervised human mesh recovery has been made using synthetic-data-driven training paradigms where the model is trained from synthetic paired 2D representation (e.g., 2D keypoints and segmentation masks) and 3D mesh. However, on synthetic dense correspondence maps (i.e., IUV) few have been explored since the domain gap between synthetic training data and real testing data is hard to address for 2D dense representation. To alleviate this domain gap on IUV, we propose cross-representation alignment utilizing the complementary information from the robust but sparse representation (2D keypoints). Specifically, the alignment errors between initial mesh estimation and both 2D representations are forwarded into regressor and dynamically corrected in the following mesh regression. This adaptive cross-representation alignment explicitly learns from the deviations and captures complementary information: robustness from sparse representation and richness from dense representation. We conduct extensive experiments on multiple standard benchmark datasets and demonstrate competitive results, helping take a step towards reducing the annotation effort needed to produce state-of-the-art models in human mesh estimation.KeywordsHuman mesh recoveryRepresentation alignmentSynthetic-to-real learning" @default.
- W4313003742 created "2023-01-05" @default.
- W4313003742 creator A5003798053 @default.
- W4313003742 creator A5003875781 @default.
- W4313003742 creator A5038526207 @default.
- W4313003742 creator A5038942290 @default.
- W4313003742 creator A5039708825 @default.
- W4313003742 creator A5044410531 @default.
- W4313003742 creator A5068780385 @default.
- W4313003742 date "2022-01-01" @default.
- W4313003742 modified "2023-09-26" @default.
- W4313003742 title "Self-supervised Human Mesh Recovery with Cross-Representation Alignment" @default.
- W4313003742 cites W1967554269 @default.
- W4313003742 cites W1989191365 @default.
- W4313003742 cites W2101032778 @default.
- W4313003742 cites W2122633688 @default.
- W4313003742 cites W2194775991 @default.
- W4313003742 cites W2483862638 @default.
- W4313003742 cites W2573098616 @default.
- W4313003742 cites W2576289912 @default.
- W4313003742 cites W2797184202 @default.
- W4313003742 cites W2798637590 @default.
- W4313003742 cites W2798646183 @default.
- W4313003742 cites W2895748257 @default.
- W4313003742 cites W2949924544 @default.
- W4313003742 cites W2962754033 @default.
- W4313003742 cites W2963150697 @default.
- W4313003742 cites W2963441822 @default.
- W4313003742 cites W2963475767 @default.
- W4313003742 cites W2963876278 @default.
- W4313003742 cites W2963907666 @default.
- W4313003742 cites W2963995996 @default.
- W4313003742 cites W2965523038 @default.
- W4313003742 cites W2968940310 @default.
- W4313003742 cites W2978956737 @default.
- W4313003742 cites W2981637078 @default.
- W4313003742 cites W2991621301 @default.
- W4313003742 cites W2994220412 @default.
- W4313003742 cites W2997288107 @default.
- W4313003742 cites W3004162361 @default.
- W4313003742 cites W3021998000 @default.
- W4313003742 cites W3035072447 @default.
- W4313003742 cites W3035501466 @default.
- W4313003742 cites W3035551320 @default.
- W4313003742 cites W3035581100 @default.
- W4313003742 cites W3096693210 @default.
- W4313003742 cites W3106758349 @default.
- W4313003742 cites W3107073427 @default.
- W4313003742 cites W3107346586 @default.
- W4313003742 cites W3107880069 @default.
- W4313003742 cites W3135266094 @default.
- W4313003742 cites W3163554922 @default.
- W4313003742 cites W3167491448 @default.
- W4313003742 cites W3167542203 @default.
- W4313003742 cites W3171097378 @default.
- W4313003742 cites W3175199633 @default.
- W4313003742 cites W3176070262 @default.
- W4313003742 cites W3177864674 @default.
- W4313003742 cites W3177949351 @default.
- W4313003742 cites W3202716970 @default.
- W4313003742 cites W3203030863 @default.
- W4313003742 cites W3203211072 @default.
- W4313003742 cites W3203617912 @default.
- W4313003742 cites W3203745449 @default.
- W4313003742 cites W3204956438 @default.
- W4313003742 cites W4206210425 @default.
- W4313003742 cites W4214517305 @default.
- W4313003742 cites W4214770715 @default.
- W4313003742 cites W4214821076 @default.
- W4313003742 cites W4231018136 @default.
- W4313003742 cites W4251108253 @default.
- W4313003742 cites W4296053296 @default.
- W4313003742 doi "https://doi.org/10.1007/978-3-031-19769-7_13" @default.
- W4313003742 hasPublicationYear "2022" @default.
- W4313003742 type Work @default.
- W4313003742 citedByCount "0" @default.
- W4313003742 crossrefType "book-chapter" @default.
- W4313003742 hasAuthorship W4313003742A5003798053 @default.
- W4313003742 hasAuthorship W4313003742A5003875781 @default.
- W4313003742 hasAuthorship W4313003742A5038526207 @default.
- W4313003742 hasAuthorship W4313003742A5038942290 @default.
- W4313003742 hasAuthorship W4313003742A5039708825 @default.
- W4313003742 hasAuthorship W4313003742A5044410531 @default.
- W4313003742 hasAuthorship W4313003742A5068780385 @default.
- W4313003742 hasBestOaLocation W43130037422 @default.
- W4313003742 hasConcept C104317684 @default.
- W4313003742 hasConcept C105795698 @default.
- W4313003742 hasConcept C119857082 @default.
- W4313003742 hasConcept C124066611 @default.
- W4313003742 hasConcept C124101348 @default.
- W4313003742 hasConcept C13280743 @default.
- W4313003742 hasConcept C153180895 @default.
- W4313003742 hasConcept C154945302 @default.
- W4313003742 hasConcept C160920958 @default.
- W4313003742 hasConcept C17744445 @default.
- W4313003742 hasConcept C185592680 @default.
- W4313003742 hasConcept C185798385 @default.
- W4313003742 hasConcept C199539241 @default.