Matches in SemOpenAlex for { <https://semopenalex.org/work/W4225557002> ?p ?o ?g. }
- W4225557002 endingPage "1293" @default.
- W4225557002 startingPage "1282" @default.
- W4225557002 abstract "Despite the great progress in 3D human pose estimation from videos, it is still an open problem to take full advantage of a redundant 2D pose sequence to learn representative representations for generating one 3D pose. To this end, we propose an improved Transformer-based architecture, called Strided Transformer, which simply and effectively lifts a long sequence of 2D joint locations to a single 3D pose. Specifically, a Vanilla Transformer Encoder (VTE) is adopted to model long-range dependencies of 2D pose sequences. To reduce the redundancy of the sequence, fully-connected layers in the feed-forward network of VTE are replaced with strided convolutions to progressively shrink the sequence length and aggregate information from local contexts. The modified VTE is termed as Strided Transformer Encoder (STE), which is built upon the outputs of VTE. STE not only effectively aggregates long-range information to a single-vector representation in a hierarchical global and local fashion, but also significantly reduces the computation cost. Furthermore, a full-to-single supervision scheme is designed at both full sequence and single target frame scales applied to the outputs of VTE and STE, respectively. This scheme imposes extra temporal smoothness constraints in conjunction with the single target frame supervision and hence helps produce smoother and more accurate 3D poses. The proposed Strided Transformer is evaluated on two challenging benchmark datasets, Human3.6 M and HumanEva-I, and achieves state-of-the-art results with fewer parameters. Code and models are available at <uri xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>https://github.com/Vegetebird/StridedTransformer-Pose3D</uri> ." @default.
- W4225557002 created "2022-05-05" @default.
- W4225557002 creator A5015140853 @default.
- W4225557002 creator A5028070258 @default.
- W4225557002 creator A5038044561 @default.
- W4225557002 creator A5042680345 @default.
- W4225557002 creator A5051601236 @default.
- W4225557002 creator A5057287134 @default.
- W4225557002 date "2023-01-01" @default.
- W4225557002 modified "2023-09-30" @default.
- W4225557002 title "Exploiting Temporal Contexts With Strided Transformer for 3D Human Pose Estimation" @default.
- W4225557002 cites W2099333815 @default.
- W4225557002 cites W2101032778 @default.
- W4225557002 cites W2134704262 @default.
- W4225557002 cites W2293220651 @default.
- W4225557002 cites W2307770531 @default.
- W4225557002 cites W2554247908 @default.
- W4225557002 cites W2612706635 @default.
- W4225557002 cites W2650426111 @default.
- W4225557002 cites W2761406536 @default.
- W4225557002 cites W2798646183 @default.
- W4225557002 cites W2799211965 @default.
- W4225557002 cites W2805088284 @default.
- W4225557002 cites W2809890486 @default.
- W4225557002 cites W2886552305 @default.
- W4225557002 cites W2895689136 @default.
- W4225557002 cites W2908684875 @default.
- W4225557002 cites W2913397815 @default.
- W4225557002 cites W2962722475 @default.
- W4225557002 cites W2962824791 @default.
- W4225557002 cites W2962896489 @default.
- W4225557002 cites W2963076553 @default.
- W4225557002 cites W2963150697 @default.
- W4225557002 cites W2963225971 @default.
- W4225557002 cites W2963441822 @default.
- W4225557002 cites W2963598138 @default.
- W4225557002 cites W2964016027 @default.
- W4225557002 cites W2964221239 @default.
- W4225557002 cites W2969450957 @default.
- W4225557002 cites W2972662547 @default.
- W4225557002 cites W2981660954 @default.
- W4225557002 cites W2984280075 @default.
- W4225557002 cites W3034448411 @default.
- W4225557002 cites W3034581612 @default.
- W4225557002 cites W3092041969 @default.
- W4225557002 cites W3092900809 @default.
- W4225557002 cites W3096609285 @default.
- W4225557002 cites W3097623574 @default.
- W4225557002 cites W3098473649 @default.
- W4225557002 cites W3098612954 @default.
- W4225557002 cites W3106838237 @default.
- W4225557002 cites W3106882556 @default.
- W4225557002 cites W3118589616 @default.
- W4225557002 cites W3118802333 @default.
- W4225557002 cites W3121523901 @default.
- W4225557002 cites W3126541466 @default.
- W4225557002 cites W3159460161 @default.
- W4225557002 cites W3169891778 @default.
- W4225557002 cites W3173811519 @default.
- W4225557002 cites W3175199633 @default.
- W4225557002 cites W3207613026 @default.
- W4225557002 cites W4214736485 @default.
- W4225557002 cites W4248675569 @default.
- W4225557002 doi "https://doi.org/10.1109/tmm.2022.3141231" @default.
- W4225557002 hasPublicationYear "2023" @default.
- W4225557002 type Work @default.
- W4225557002 citedByCount "30" @default.
- W4225557002 countsByYear W42255570022022 @default.
- W4225557002 countsByYear W42255570022023 @default.
- W4225557002 crossrefType "journal-article" @default.
- W4225557002 hasAuthorship W4225557002A5015140853 @default.
- W4225557002 hasAuthorship W4225557002A5028070258 @default.
- W4225557002 hasAuthorship W4225557002A5038044561 @default.
- W4225557002 hasAuthorship W4225557002A5042680345 @default.
- W4225557002 hasAuthorship W4225557002A5051601236 @default.
- W4225557002 hasAuthorship W4225557002A5057287134 @default.
- W4225557002 hasBestOaLocation W42255570022 @default.
- W4225557002 hasConcept C111919701 @default.
- W4225557002 hasConcept C11413529 @default.
- W4225557002 hasConcept C118505674 @default.
- W4225557002 hasConcept C121332964 @default.
- W4225557002 hasConcept C152124472 @default.
- W4225557002 hasConcept C153180895 @default.
- W4225557002 hasConcept C154945302 @default.
- W4225557002 hasConcept C165801399 @default.
- W4225557002 hasConcept C41008148 @default.
- W4225557002 hasConcept C45374587 @default.
- W4225557002 hasConcept C52102323 @default.
- W4225557002 hasConcept C62520636 @default.
- W4225557002 hasConcept C66322947 @default.
- W4225557002 hasConceptScore W4225557002C111919701 @default.
- W4225557002 hasConceptScore W4225557002C11413529 @default.
- W4225557002 hasConceptScore W4225557002C118505674 @default.
- W4225557002 hasConceptScore W4225557002C121332964 @default.
- W4225557002 hasConceptScore W4225557002C152124472 @default.
- W4225557002 hasConceptScore W4225557002C153180895 @default.
- W4225557002 hasConceptScore W4225557002C154945302 @default.
- W4225557002 hasConceptScore W4225557002C165801399 @default.