Lune

ICCV2023Top-tier venue

SEFD: Learning to Distill Complex Pose and Occlusion

ChangHee Yang, Kyeongbo Kong, Sung-Jun Min, Dongyoon Wee, Ho-Deok Jang, Geonho Cha, Suk-Ju Kang

Abstract

This paper addresses the problem of three-dimensional (3D) human mesh estimation in complex poses and occluded situations. Although many improvements have been made in 3D human mesh estimation using the twodimensional (2D) pose with occlusion between humans, occlusion from complex poses and other objects remains a consistent problem. Therefore, we propose the novel Skinned Multi-Person Linear (SMPL) Edge Feature Distillation (SEFD) that demonstrates robustness to complex poses and occlusions, without increasing the number of parameters compared to the baseline model. The model generates an SMPL overlapping edge similar to the ground truth that contains target person boundary and occlusion information, performing subsequent feature distillation in a simple edge map. We also perform experiments on various benchmarks and exhibit fidelity both qualitatively and quantitatively. Extensive experiments prove that our method outperforms the state-of-the-art method by 2.8% in MPJPE and 1.9% in MPVPE on a benchmark 3DPW dataset in the presence of domain gap. Also, our method is superior in 3DPW-OCC, 3DPW-PC, RH-Dataset, OCHuman, Crowd-Pose, and LSP dataset in which occlusion, complex pose, and domain gap exist. The code and occlusion & complex pose annotation will be available at https://github . com/YangChangHee/ICCV2023_SEFD_RELEASE/.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

2 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Mesh Graphormer cites Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the Loop
  • PoseSyn: Synthesizing Diverse 3D Pose Data from In-the-Wild 2D Data cites Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the Loop
  • PARE: Part Attention Regressor for 3D Human Body Estimation cites Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the Loop
  • MEGA: Masked Generative Autoencoder for Human Mesh Recovery cites Mesh Graphormer
  • MEGA: Masked Generative Autoencoder for Human Mesh Recovery cites Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the Loop
  • MEGA: Masked Generative Autoencoder for Human Mesh Recovery cites PARE: Part Attention Regressor for 3D Human Body Estimation
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers2

  1. PoseSyn: Synthesizing Diverse 3D Pose Data from In-the-Wild 2D DataChangHee Yang, Hyeonseop Song, Seokhun Choi, Seungwoo Lee et al.ICCV 2025 · 1 citation
  2. MEGA: Masked Generative Autoencoder for Human Mesh RecoveryGuénolé Fiche, Simon Leglaive, Xavier Alameda-Pineda, Francesc Moreno-NoguerCVPR 2025

Builds on27

  1. Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the LoopNikos Kolotouros, Georgios Pavlakos, Michael J. Black, Kostas DaniilidisICCV 2019 · 1,139 citations
  2. EGNet: Edge Guidance Network for Salient Object DetectionJiaxing Zhao, Jiang-Jiang Liu, Deng-Ping Fan, Yang Cao et al.ICCV 2019 · 1,054 citations
  3. PARE: Part Attention Regressor for 3D Human Body EstimationMuhammed Kocabas, Chun-Hao P. Huang, Otmar Hilliges, Michael J. BlackICCV 2021 · 509 citations
  4. Pixel Difference Networks for Efficient Edge DetectionZhuo Su, Wenzhe Liu, Zitong Yu, Dewen Hu et al.ICCV 2021 · 488 citations
  5. Mesh GraphormerKevin Lin, Lijuan Wang, Zicheng LiuICCV 2021 · 399 citations

Related papers

  1. Learning to Estimate Robust 3D Human Mesh from In-the-Wild Crowded ScenesHongsuk Choi, Gyeongsik Moon, JoonKyu Park, Kyoung Mu LeeCVPR 2022 · 92 citations
  2. UniSH: Unifying Scene and Human Reconstruction in a Feed-Forward PassMengfei Li, Peng Li, Zheng Zhang, Jiahao Lu et al.CVPR 2026 · 7 citations
  3. Pose-Guided Feature Learning with Knowledge Distillation for Occluded Person Re-IdentificationKecheng Zheng, Cuiling Lan, Wenjun Zeng, Jiawei Liu et al.ACM MM 2021 · 81 citations
  4. Probabilistic 3D Human Shape and Pose Estimation From Multiple Unconstrained Images in the WildAkash Sengupta, Ignas Budvytis, Roberto CipollaCVPR 2021
  5. Instance-Aware Contrastive Learning for Occluded Human Mesh ReconstructionMi-Gyeong Gwon, Gi-Mun Um, Won-Sik Cheong, Wonjun KimCVPR 2024
Dusk over the sea between two cliffs drawn in fine vertical lines