Pose Estimation on 3DPW

Metric: MPVPE (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	MPVPE▼	Extra Data	Paper	Date↕	Code
1	BMP	119.3	No	Body Meshes as Points	2021-05-06	Code
2	SPIN	116.4	No	Learning to Reconstruct 3D Human Pose and Shape ...	2019-09-27	Code
3	Wenshuo et a;.	112.6	No	Decomposed Human Motion Prior for Video Pose Est...	2023-05-30	-
4	TCMR (T=16 w/o 3DPW)	111.5	No	Beyond Static Features for Temporally Consistent...	2020-11-17	Code
5	CHOMP	110.1	No	Occluded Human Body Capture with Self-Supervised...	2022-07-12	Code
6	PC-HMR	108.6	No	PC-HMR: Pose Calibration for 3D Human Mesh Recov...	2021-03-16	-
7	3DCrowdNet	108.5	No	Learning to Estimate Robust 3D Human Mesh from I...	2021-04-15	Code
8	GTRS	106.2	No	A Lightweight Graph Transformer Network for Huma...	2021-11-24	Code
9	MUG	106.2	No	MUG: Multi-human Graph Network for 3D Mesh Recon...	2022-05-25	-
10	MoCap-SPIN + PoseBERT	103.8	No	Leveraging MoCap Data for Human Mesh Recovery	2021-10-18	Code
11	TCMR (T=16 w/o H3.6M)	102.9	No	Beyond Static Features for Temporally Consistent...	2020-11-17	Code
12	CLIFF	102.1	No	3D Human Reconstruction in the Wild with Synthet...	2024-03-17	-
13	STR	101.2	No	Spatio-temporal Tendency Reasoning for Human Bod...	2022-10-07	-
14	DDT	101.2	No	DiffMesh: A Motion-aware Diffusion Framework for...	2023-03-23	-
15	Self-Attentive	100.6	No	Self-Attentive 3D Human Pose and Shape Estimatio...	2021-03-26	-
16	TePose (T=6)	100.3	No	Live Stream Temporally Embedded 3D Human Body Po...	2022-07-25	Code
17	CycleAdapt (w/o 2D GT)	99.9	No	Cyclic Test-Time Adaptation on Monocular Video f...	2023-08-12	Code
18	MPS-Net (T=16)	99.7	No	Capturing Humans in Motion: Temporal-Attentive 3...	2022-03-16	-
19	VIBE	99.1	No	VIBE: Video Inference for Human Body Pose and Sh...	2019-12-11	Code
20	TAPE (T=16)	98.1	No	TAPE: Temporal Attention-based Probabilistic hum...	2023-04-29	Code
21	TRACE	97.3	No	TRACE: 5D Temporal Regression of Avatars with Dy...	2023-06-05	Code
22	KAMA	97	No	KAMA: 3D Keypoint Aware Body Mesh Articulation	2021-04-27	-
23	GLoT	96.3	No	Global-to-Local Modeling for Video-based 3D Huma...	2023-03-26	Code
24	LMR	93.6	No	Learning Local Recurrent Models for Human Mesh R...	2021-07-27	-
25	DST-VIBE	93.5	No	Deep Two-Stream Video Inference for Human Body P...	2021-10-22	-
26	SEFD	92.6	No	-	-	Code
27	BOA (w/ 2D GT)	91.2	No	Bilevel Online Adaptation for Out-of-Domain Huma...	2021-03-30	Code
28	MotionBERT (Finetune)	88.1	No	MotionBERT: A Unified Perspective on Learning Hu...	2022-10-12	Code
29	INT-2 (ResNet-50)	87.9	No	Capturing the motion of every joint: 3D human po...	2023-03-01	Code
30	POTTER	87.4	No	POTTER: Pooling Attention Transformer for Effici...	2023-03-23	Code
31	ImpHMR	87.1	No	Implicit 3D Human Mesh Recovery using Consistenc...	2023-06-30	-
32	XFormer (HRNet)	87.1	No	XFormer: Fast and Accurate Monocular 3D Body Cap...	2023-05-18	-
33	HeatER	86.9	No	FeatER: An Efficient Network for Human Reconstru...	2022-05-30	Code
34	NIKI (Twist-and-Swing)	86.6	No	NIKI: Neural Inverse Kinematics with Invertible ...	2023-05-15	Code
35	HybrIK	86.5	No	HybrIK: A Hybrid Analytical-Neural Inverse Kinem...	2020-11-30	Code
36	TokenHMR (SD + ITW + BL)	84.6	No	TokenHMR: Advancing Human Mesh Recovery with a T...	2024-04-25	Code
37	IKOL	84.1	No	IKOL: Inverse kinematics optimization layer for ...	2023-02-02	Code
38	PSVT	84	No	PSVT: End-to-End Multi-person 3D Pose and Shape ...	2023-03-16	-
39	HybrIK-Transformer (HrNet-48)	83.6	No	3D Human Pose and Shape Estimation via HybrIK-Tr...	2023-02-09	Code
40	DeFormer	82.6	No	-	-	Code
41	HybrIK (HRNet-W48)	82.3	No	HybrIK-X: Hybrid Analytical-Neural Inverse Kinem...	2023-04-12	Code
42	HMR 2.0	82.2	No	Humans in 4D: Reconstructing and Tracking Humans...	2023-05-31	Code
43	DynaBOA (w/ 2D GT)	82	No	Out-of-Domain Human Mesh Reconstruction via Dyna...	2021-11-07	Code
44	SMPLer-L	82	No	SMPLer: Taming Transformers for Monocular 3D Hum...	2024-04-23	Code
45	RemoCap	81.9	No	RemoCap: Disentangled Representation Learning fo...	2024-05-21	-
46	ARTS (Resnet50 L=16)	81.4	No	ARTS: Semi-Analytical Regressor using Disentangl...	2024-10-21	Code
47	BoPR (HR-W48)	80.8	No	BoPR: Body-aware Part Regressor for Human Shape ...	2023-03-21	Code
48	BioPose	79.8	No	BioPose: Biomechanically-accurate 3D Pose Estima...	2025-01-14	-
49	Hulk(ViT-B)	79.8	No	Hulk: A Universal Knowledge Translator for Human...	2023-12-04	Code
50	MotionBERT-HybrIK	79.4	No	MotionBERT: A Unified Perspective on Learning Hu...	2022-10-12	Code
51	MPT	79.4	No	MPT: Mesh Pre-Training with Transformers for Hum...	2022-11-24	Code
52	PROM (CLIFF)	79.2	No	Learning Unorthogonalized Matrices for Rotation ...	2023-12-01	-
53	SEFD_GT	78.36	No	-	-	Code
54	PostoMETRO (ResNet-50)	78	No	PostoMETRO: Pose Token Enhanced Mesh Transformer...	2024-03-19	-
55	VirtualMarker	77.9	No	3D Human Mesh Estimation from Virtual Markers	2023-03-21	Code
56	GenHMR	77.5	No	GenHMR: Generative Human Mesh Recovery	2024-12-19	-
57	Hulk(ViT-L)	77.4	No	Hulk: A Universal Knowledge Translator for Human...	2023-12-04	Code
58	PostoMETRO (HRNet-w48)	76.8	No	PostoMETRO: Pose Token Enhanced Mesh Transformer...	2024-03-19	-
59	CLIFF (3DPW+HumanWild+BEDLAM+AGORA)	76.8	No	3D Human Reconstruction in the Wild with Synthet...	2024-03-17	-
60	CycleAdapt (w/ 2D GT)	76.7	No	Cyclic Test-Time Adaptation on Monocular Video f...	2023-08-12	Code
61	Zolly (HRNet-w48)	76.3	No	Zolly: Zoom Focal Length Correctly for Perspecti...	2023-03-24	Code
62	TAR (N=9)	74.4	No	Temporal-Aware Refinement for Video-based Human ...	2023-11-16	-
63	WHAM (ViT)	68.7	No	WHAM: Reconstructing World-grounded Humans with ...	2023-12-12	Code
64	ExtPose (-a, T=16)	67.5	No	-	-	-

#1BMPSOTA
119.3
MPVPE· 2021-05-06
Body Meshes as Points Code
#2SPINSOTA
116.4
MPVPE· 2019-09-27
Learning to Reconstruct 3D Human Pose and Shape via Model-fitting in the Loop Code
#3Wenshuo et a;.
112.6
MPVPE· 2023-05-30
Decomposed Human Motion Prior for Video Pose Estimation via Adversarial Training
#4TCMR (T=16 w/o 3DPW)
111.5
MPVPE· 2020-11-17
Beyond Static Features for Temporally Consistent 3D Human Pose and Shape from a Video Code
#5CHOMP
110.1
MPVPE· 2022-07-12
Occluded Human Body Capture with Self-Supervised Spatial-Temporal Motion Prior Code
#6PC-HMR
108.6
MPVPE· 2021-03-16
PC-HMR: Pose Calibration for 3D Human Mesh Recovery from 2D Images/Videos
#73DCrowdNet
108.5
MPVPE· 2021-04-15
Learning to Estimate Robust 3D Human Mesh from In-the-Wild Crowded Scenes Code
#8GTRS
106.2
MPVPE· 2021-11-24
A Lightweight Graph Transformer Network for Human Mesh Reconstruction from 2D Human Pose Code
#9MUG
106.2
MPVPE· 2022-05-25
MUG: Multi-human Graph Network for 3D Mesh Reconstruction from 2D Pose
#10MoCap-SPIN + PoseBERT
103.8
MPVPE· 2021-10-18
Leveraging MoCap Data for Human Mesh Recovery Code
#11TCMR (T=16 w/o H3.6M)
102.9
MPVPE· 2020-11-17
Beyond Static Features for Temporally Consistent 3D Human Pose and Shape from a Video Code
#12CLIFF
102.1
MPVPE· 2024-03-17
3D Human Reconstruction in the Wild with Synthetic Data Using Generative Models
#13STR
101.2
MPVPE· 2022-10-07
Spatio-temporal Tendency Reasoning for Human Body Pose and Shape Estimation from Videos
#14DDT
101.2
MPVPE· 2023-03-23
DiffMesh: A Motion-aware Diffusion Framework for Human Mesh Recovery from Videos
#15Self-Attentive
100.6
MPVPE· 2021-03-26
Self-Attentive 3D Human Pose and Shape Estimation from Videos
#16TePose (T=6)
100.3
MPVPE· 2022-07-25
Live Stream Temporally Embedded 3D Human Body Pose and Shape Estimation Code
#17CycleAdapt (w/o 2D GT)
99.9
MPVPE· 2023-08-12
Cyclic Test-Time Adaptation on Monocular Video for 3D Human Mesh Reconstruction Code
#18MPS-Net (T=16)
99.7
MPVPE· 2022-03-16
Capturing Humans in Motion: Temporal-Attentive 3D Human Pose and Shape Estimation from Monocular Video
#19VIBE
99.1
MPVPE· 2019-12-11
VIBE: Video Inference for Human Body Pose and Shape Estimation Code
#20TAPE (T=16)
98.1
MPVPE· 2023-04-29
TAPE: Temporal Attention-based Probabilistic human pose and shape Estimation Code
#21TRACE
97.3
MPVPE· 2023-06-05
TRACE: 5D Temporal Regression of Avatars with Dynamic Cameras in 3D Environments Code
#22KAMA
97
MPVPE· 2021-04-27
KAMA: 3D Keypoint Aware Body Mesh Articulation
#23GLoT
96.3
MPVPE· 2023-03-26
Global-to-Local Modeling for Video-based 3D Human Pose and Shape Estimation Code
#24LMR
93.6
MPVPE· 2021-07-27
Learning Local Recurrent Models for Human Mesh Recovery
#25DST-VIBE
93.5
MPVPE· 2021-10-22
Deep Two-Stream Video Inference for Human Body Pose and Shape Estimation
#26SEFD
92.6
MPVPE
No paperCode
#27BOA (w/ 2D GT)
91.2
MPVPE· 2021-03-30
Bilevel Online Adaptation for Out-of-Domain Human Mesh Reconstruction Code
#28MotionBERT (Finetune)
88.1
MPVPE· 2022-10-12
MotionBERT: A Unified Perspective on Learning Human Motion Representations Code
#29INT-2 (ResNet-50)
87.9
MPVPE· 2023-03-01
Capturing the motion of every joint: 3D human pose and shape estimation with independent tokens Code
#30POTTER
87.4
MPVPE· 2023-03-23
POTTER: Pooling Attention Transformer for Efficient Human Mesh Recovery Code
#31ImpHMR
87.1
MPVPE· 2023-06-30
Implicit 3D Human Mesh Recovery using Consistency with Pose and Shape from Unseen-view
#32XFormer (HRNet)
87.1
MPVPE· 2023-05-18
XFormer: Fast and Accurate Monocular 3D Body Capture
#33HeatER
86.9
MPVPE· 2022-05-30
FeatER: An Efficient Network for Human Reconstruction via Feature Map-Based TransformER Code
#34NIKI (Twist-and-Swing)
86.6
MPVPE· 2023-05-15
NIKI: Neural Inverse Kinematics with Invertible Neural Networks for 3D Human Pose and Shape Estimation Code
#35HybrIK
86.5
MPVPE· 2020-11-30
HybrIK: A Hybrid Analytical-Neural Inverse Kinematics Solution for 3D Human Pose and Shape Estimation Code
#36TokenHMR (SD + ITW + BL)
84.6
MPVPE· 2024-04-25
TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose Representation Code
#37IKOL
84.1
MPVPE· 2023-02-02
IKOL: Inverse kinematics optimization layer for 3D human pose and shape estimation via Gauss-Newton differentiation Code
#38PSVT
84
MPVPE· 2023-03-16
PSVT: End-to-End Multi-person 3D Pose and Shape Estimation with Progressive Video Transformers
#39HybrIK-Transformer (HrNet-48)
83.6
MPVPE· 2023-02-09
3D Human Pose and Shape Estimation via HybrIK-Transformer Code
#40DeFormer
82.6
MPVPE
No paperCode
#41HybrIK (HRNet-W48)
82.3
MPVPE· 2023-04-12
HybrIK-X: Hybrid Analytical-Neural Inverse Kinematics for Whole-body Mesh Recovery Code
#42HMR 2.0
82.2
MPVPE· 2023-05-31
Humans in 4D: Reconstructing and Tracking Humans with Transformers Code
#43DynaBOA (w/ 2D GT)
82
MPVPE· 2021-11-07
Out-of-Domain Human Mesh Reconstruction via Dynamic Bilevel Online Adaptation Code
#44SMPLer-L
82
MPVPE· 2024-04-23
SMPLer: Taming Transformers for Monocular 3D Human Shape and Pose Estimation Code
#45RemoCap
81.9
MPVPE· 2024-05-21
RemoCap: Disentangled Representation Learning for Motion Capture
#46ARTS (Resnet50 L=16)
81.4
MPVPE· 2024-10-21
ARTS: Semi-Analytical Regressor using Disentangled Skeletal Representations for Human Mesh Recovery from Videos Code
#47BoPR (HR-W48)
80.8
MPVPE· 2023-03-21
BoPR: Body-aware Part Regressor for Human Shape and Pose Estimation Code
#48BioPose
79.8
MPVPE· 2025-01-14
BioPose: Biomechanically-accurate 3D Pose Estimation from Monocular Videos
#49Hulk(ViT-B)
79.8
MPVPE· 2023-12-04
Hulk: A Universal Knowledge Translator for Human-Centric Tasks Code
#50MotionBERT-HybrIK
79.4
MPVPE· 2022-10-12
MotionBERT: A Unified Perspective on Learning Human Motion Representations Code
#51MPT
79.4
MPVPE· 2022-11-24
MPT: Mesh Pre-Training with Transformers for Human Pose and Mesh Reconstruction Code
#52PROM (CLIFF)
79.2
MPVPE· 2023-12-01
Learning Unorthogonalized Matrices for Rotation Estimation
#53SEFD_GT
78.36
MPVPE
No paperCode
#54PostoMETRO (ResNet-50)
78
MPVPE· 2024-03-19
PostoMETRO: Pose Token Enhanced Mesh Transformer for Robust 3D Human Mesh Recovery
#55VirtualMarker
77.9
MPVPE· 2023-03-21
3D Human Mesh Estimation from Virtual Markers Code
#56GenHMR
77.5
MPVPE· 2024-12-19
GenHMR: Generative Human Mesh Recovery
#57Hulk(ViT-L)
77.4
MPVPE· 2023-12-04
Hulk: A Universal Knowledge Translator for Human-Centric Tasks Code
#58PostoMETRO (HRNet-w48)
76.8
MPVPE· 2024-03-19
PostoMETRO: Pose Token Enhanced Mesh Transformer for Robust 3D Human Mesh Recovery
#59CLIFF (3DPW+HumanWild+BEDLAM+AGORA)
76.8
MPVPE· 2024-03-17
3D Human Reconstruction in the Wild with Synthetic Data Using Generative Models
#60CycleAdapt (w/ 2D GT)
76.7
MPVPE· 2023-08-12
Cyclic Test-Time Adaptation on Monocular Video for 3D Human Mesh Reconstruction Code
#61Zolly (HRNet-w48)
76.3
MPVPE· 2023-03-24
Zolly: Zoom Focal Length Correctly for Perspective-Distorted Human Mesh Reconstruction Code
#62TAR (N=9)
74.4
MPVPE· 2023-11-16
Temporal-Aware Refinement for Video-based Human Pose and Shape Recovery
#63WHAM (ViT)
68.7
MPVPE· 2023-12-12
WHAM: Reconstructing World-grounded Humans with Accurate 3D Motion Code
#64ExtPose (-a, T=16)
67.5
MPVPE
No paper