We introduce SAM 3D Body (3DB), a promptable model for single-image full-body 3D human mesh recovery (HMR) that demonstrates state-of-the-art performance, with strong generalization and consistent accuracy in diverse in-the-wild conditions. 3DB estimates the human pose of the body, feet, and hands. It is the first model to use a new parametric mesh representation, Momentum Human Rig (MHR), which decouples skeletal structure and surface shape. 3DB employs an encoder-decoder architecture and supports auxiliary prompts, including 2D keypoints and masks, enabling user-guided inference similar to the SAM family of models. We derive high-quality annotations from a multi-stage annotation pipeline that uses various combinations of manual keypoint annotation, differentiable optimization, multi-view geometry, and dense keypoint detection. Our data engine efficiently selects and processes data to ensure data diversity, collecting unusual poses and rare imaging conditions. We present a new evaluation dataset organized by pose and appearance categories, enabling nuanced analysis of model behavior. Our experiments demonstrate superior generalization and substantial improvements over prior methods in both qualitative user preference studies and traditional quantitative analysis. Both 3DB and MHR are open-source.
Figures & tables
Figure 2: SAM 3D Body Model Architecture. We employ a promptable encoder–decoder architecture with a shared image encoder and separate decoders for body and hand pose estimation.
Figure 3: Left: GUI of our annotation tool for annotating 2D keypoints. Right: Comparison of the dense (thin) and sparse (thick) keypoints for pseudo annotation.
Figure 4: Example of single-image MHR mesh fitting for ITW datasets. Source: SA-1B Kirillov et al. (2023) .
Figure 5: Examples of MHR mesh fitting results. (a) Multi-view mesh fitting. Source: EgoExo4D Grauman et al. (2024) . (b) Scan-based mesh fitting. Source: Re:Interhand Moon et al. (2023) .
Table 2: Comparison on five common benchmarks. The best results are highlighted in bold, while the second-best results are underlined. Results evaluated using publicly released checkpoint denoted by † . Models trained using RICH denoted by ∗ .
EE4D-Phy
EE4D-Proc
Harmony4D
Goliath
Synthetic
SA1B-Hard
Models
PVE ↓
MPJPE ↓
PVE ↓
MPJPE ↓
PVE ↓
MPJPE ↓
PVE ↓
MPJPE ↓
PVE ↓
MPJPE ↓
Avg-PCK ↑
CameraHMR Patel and Black (2025)
71.1
58.8
70.3
60.2
84.6
70.8
66.7
54.5
102.8
87.2
63.0
PromptHMR Wang et al. (2025c)
74.6
63.4
72.0
62.6
91.9
78.0
67.2
56.5
92.7
80.7
59.0
NLF Sárándi and Pons-Moll (2024)
75.9
68.5
85.4
77.7
97.3
84.9
66.5
58.0
97.6
86.5
66.5
3DB-H Leave-one-out (Ours)
49.7
44.3
52.9
47.4
63.5
54.0
54.2
46.5
85.6
75.5
73.1
3DB-H Full dataset (Ours)
37.0
31.6
41.9
36.3
41.0
33.9
34.5
28.8
55.2
47.2
76.6
Table 3: Comparison on five new benchmark datasets. The best results are highlighted in bold, while the second-best results are underlined. MPJPE is computed on 24 SMPL keypoints.
Method
PA-MPVPE ↓
PA-MPJPE ↓
F@5 ↑
F@15 ↑
LookMa Hewitt et al. (2024)
8.1
8.6
0.653
-
METRO Lin et al. (2021) †
6.3
6.5
0.731
0.984
HaMeR Pavlakos et al. (2024) †
5.7
6.0
0.785
0.990
MaskHand Saleem et al. (2025) †
5.4
5.5
0.801
0.991
WiLoR Potamias et al. (2025) †
5.1
5.5
0.825
0.993
3DB-H (Ours)
6.3
5.5
0.735
0.988
Table 4: Comparison on Freihand for hand pose estimation. Methods using Freihand for training are denoted by † .
CameraHMR Patel and Black (2025)
PromptHMR Wang et al. (2025c)
3DB
aPCK (body)
aPCK (feet)
aPCK (body)
aPCK (feet)
aPCK (body)
aPCK (feet)
Body_shape - In-the-wild
87.64
78.56
85.73
77.87
90.76
92.12
Camera_view - Back or side view
59.69
46.64
61.92
47.74
76.27
66.81
Camera_view - Bottom-up view
55.18
34.84
46.56
29.25
69.62
55.35
Camera_view - Others
51.48
33.80
54.39
38.55
76.62
71.52
Camera_view - Overhead view
55.08
39.46
43.65
24.63
73.33
66.94
Table 5: 2D categorical performance analysis on the SA-1B Hard dataset.
CameraHMR Patel and Black (2025)
PromptHMR Wang et al. (2025c)
3DB
PVE
MPJPE
PA-MPJPE
PVE
MPJPE
PA-MPJPE
PVE
MPJPE
PA-MPJPE
aux:depth_ambiguous
126.25
102.25
81.33
109.58
91.77
69.24
64.38
52.72
39.85
aux:orient_ambiguous
84.26
71.77
45.07
83.79
72.93
46.17
42.35
36.64
25.16
aux:scale_ambiguous
118.18
104.77
50.93
112.95
102.28
47.26
58.64
51.16
27.67
fov:medium
82.88
68.81
46.86
76.31
64.84
42.85
43.58
36.97
25.57
fov:narrow
82.15
69.82
49.73
90.41
77.95
53.49
52.14
43.89
36.18
Table 6: 3D categorical performance analysis.
Figure 6: Qualitative comparison of 3DB against state-of-the-art HMR methods. Source: SA-1B Kirillov et al. (2023) .
Figure 7: Qualitative results of hand estimation using the hand decoder of 3DB. Source: Freihand Zimmermann et al. (2019) .
Figure 8: Comparison of 3DB win rate against baselines for human preference study. Win rate (%) and number of wins out of 80.
Table 7: Ablation on 2D keypoint prompting with 3DB-H. We report results under varying numbers of prompts, as well as different noise scales for a single prompt.
Hi4D Yin et al. [2023]
Harmony4D
SA1B-Hard
SA1B-MP
Models
PVE ↓
MPJPE ↓
PVE ↓
MPJPE ↓
Avg-PCK ↑
Avg-PCK ↑
3DB (w/o mask)
91.4
76.4
42.7
35.6
75.4
67.9
3DB (w/ mask)
58.3
47.0
36.5
30.1
76.3
72.3
Appendix
Table 8: Comparison on mask-conditioned inference with 3DB-DINOv3 on multi-person datasets.
Figure 9: Qualitative comparison to show the impact from using keypoint prompting and unifying the predictions from hand decoder and body decoder.