Self-supervised learning relies on asymmetry between what a model observes and what it must explain. Recent methods learn 3D geometry from unposed images by predicting cameras and scene structure, then rendering held-out views. To avoid trivial solutions, they impose this asymmetry in the input space, restricting how views interact through separate branches or masked attention. We revisit this design with geometric accuracy, rather than rendering quality, as the main objective. When a held-out view lies between context views, the context already constrains the geometry it observes. We therefore impose the asymmetry in the 3D representation space, excluding held-out views only from the reconstructed scene while a shared backbone uses all views for correspondence. This improves depth and pose estimation over input-space designs. We further exploit view density: sparse and dense observations depict the same scene but provide different amounts of correspondence. Aligning sparse-context features with dense-context features supplies a geometric training signal alongside rendering, without a pretrained geometric teacher. We additionally align features with a frozen vision foundation model, which brings further gains. None of these signals requires geometric annotation, and the resulting model improves geometry estimation over prior self-supervised methods. Controlled experiments isolate the contribution of each component, and self-supervised pretraining also improves subsequent supervised training.
Figures & tables
Figure 1: From unposed images alone, our self-supervised model recovers camera poses and 3D geometry in a single forward pass (top), generalizes to unseen datasets, improves pose accuracy over baseline models (bottom left, middle), and provides an effective initialization for supervised fine-tuning (SSL+SFT), compared to supervised training only (bottom right).
Figure 2: All designs build the 3D scene from context views (C). (a) Separate backbones and (b) masked attention restrict interaction between context and held-out views (H). (c) Our design allows full cross-view attention while excluding held-out views from the 3D representation.
Figure 3: The context views alone determine the same scene as all captured views. All views are shown in blue, the context subset in orange, and held-out targets in green. Held-out views add correspondence, not new geometry.
Figure 4: Effect of full interaction on depth and pose estimation. Depth is compared with MoGe-2 ( Wang et al., 2025c ) pseudo ground truth, and pose with dataset cameras.
Figure 5: Effect of asymmetric feature alignment on DL3DV dataset. The left four plots evaluate multi-view depth and camera pose; the right two evaluate depth from a single input image.
Method
3D prior
Params
RE10K (trained)
DL3DV
ScanNetv2
NRGBD
RPA@3 ↑
@5 ↑
RPA@3 ↑
@5 ↑
RPA@3 ↑
@5 ↑
AbsRel ↓
δ<1.25↑
RPA@3 ↑
@5 ↑
AbsRel ↓
δ<1.25↑
SPFSplat
MASt3R
629M
64.2
78.1
10.5
19.3
2.0
5.7
0.223
0.757
3.2
14.1
0.156
0.803
NAS3R
–
1.22B
47.4
67.2
1.5
3.5
0.1
0.5
0.381
0.604
3.7
7.2
0.214
0.630
Ours
–
183M
65.7
80.2
14.2
26.4
2.7
7.1
0.226
0.785
16.5
30.1
0.106
0.878
Table 1: Camera pose and depth accuracy with 10 input views, trained on RE10K dataset.
Figure 6: Camera pose accuracy on DL3DV dataset as the number of input views increases.
Method
Params
DL3DV (trained)
RE10K
ScanNetv2
NRGBD
RPA@3 ↑
@5 ↑
RPA@3 ↑
@5 ↑
RPA@3 ↑
@5 ↑
AbsRel ↓
δ<1.25↑
RPA@3 ↑
@5 ↑
AbsRel ↓
δ<1.25↑
E-RayZer
246M
60.3
72.1
54.6
74.5
4.9
12.3
0.168
0.819
35.6
61.5
0.095
0.908
Ours
183M
70.0
79.7
57.7
75.3
6.6
16.7
0.125
0.872
63.5
81.5
0.082
0.917
Table 2: Camera pose and depth accuracy with 10 input views, trained on DL3DV dataset.
Input
Method
ScanNetv2
NRGBD
AbsRel ↓
δ<1.25↑
AbsRel ↓
δ<1.25↑
Single image
E-RayZer
0.201
0.721
0.190
0.713
Ours
0.099
0.911
0.113
0.859
Multi-view
E-RayZer
0.168
0.819
0.095
0.908
Ours
0.125
0.872
0.082
0.917
Pose AUC@30 ↑
E-RayZer
39.1
83.1
Table 3: Single- and multi-view depth performance. Both settings are evaluated on the same indices per scene after per-view median scaling.
Figure 7: In-domain camera poses from 10 unposed views on RE10K (left) and DL3DV (right). Predictions are shown in color and ground truth in black, aligned by a similarity transform.
Figure 8: Out-of-domain camera poses on ScanNetv2 (top two rows) and NRGBD (bottom two rows), from models trained on RE10K (left) and DL3DV (right).
Method
DL3DV (trained)
ScanNetv2
NRGBD
RPA@3 ↑
@5 ↑
@10 ↑
AUC@30 ↑
RPA@3 ↑
@5 ↑
@10 ↑
AUC@30 ↑
RPA@3 ↑
@5 ↑
@10 ↑
AUC@30 ↑
SSL
66.6
81.7
91.7
88.0
1.2
5.2
17.6
29.8
16.5
37.8
74.3
72.3
SL
50.8
74.6
92.5
87.2
0.7
3.2
13.5
25.2
16.3
41.2
82.7
77.2
SSL+SFT
73.4
89.3
97.2
92.4
3.4
10.1
28.4
41.4
33.3
62.0
87.9
81.9
Table 4: Camera pose accuracy of self-supervised (SSL) and supervised (SL) training, and of self-supervised pretraining followed by supervised fine-tuning (SSL+SFT).
Method
Camera pose
Multi-view depth
Single-view depth
RPA@3 ↑
RPA@5 ↑
AbsRel ↓
δ<1.25↑
δ<1.252↑
AbsRel ↓
δ<1.25↑
δ<1.252↑
No alignment
66.6
81.7
0.195
0.758
0.901
0.230
0.638
0.850
Depth alignment
63.4
81.7
0.244
0.692
0.854
0.248
0.651
0.844
Feature alignment (ours)
69.8
84.6
0.191
0.768
0.905
0.191
0.729
0.901
Table 5: Asymmetric context alignment analysis on DL3DV dataset.
Figure 9: Single-view depth validation on DL3DV dataset over training steps. The shaded region marks the ramp-up of the alignment loss.
Training data
DL3DV
SpatialVID
benchmark
dyn_low
dyn_high
crowded
DL3DV
80.7 / 87.1
36.9 / 70.0
25.4 / 55.7
24.1 / 55.1
DL3DV + SpatialVID
79.1 / 85.8
74.7 / 86.0
73.9 / 84.0
87.4 / 89.2
Table 6: Camera pose accuracy (RPA@5 / AUC@30, higher is better) of models trained on DL3DV only and on DL3DV with SpatialVID. SpatialVID test sets are grouped by the amount of dynamic content and use a 32-frame span. DL3DV uses a 48-frame span.
Figure 10: Moving people and vehicles are largely absent from the renderings at the target views, while the static background is reconstructed.
Figure 11: Camera poses on held-out SpatialVID videos with many moving objects (left) and crowds (right), from models trained on DL3DV only and on DL3DV with SpatialVID.
Appendix figures & tables13 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 12: Overall architecture, with a shared backbone for camera and geometry prediction.
Figure 13: View synthesis evaluation. Rendering metrics on DL3DV dataset as the number of input views grows, with the context and target split shown in parentheses.
Figure 14: Depth validation over training in the posed setting.
Views
FULL
MASK
AbsRel ↓
δ<1.25↑
δ<1.252↑
PSNR ↑
LPIPS ↓
AbsRel ↓
δ<1.25↑
δ<1.252↑
PSNR ↑
LPIPS ↓
7 (4/3)
0.271 (0.274)
0.729 (0.743)
0.887 (0.892)
20.42 (21.32)
0.365 (0.328)
0.306
0.693
0.873
21.17
0.328
9 (5/4)
0.276 (0.281)
0.734 (0.746)
0.887 (0.892)
20.92 (21.74)
0.344 (0.312)
0.319
0.689
0.871
21.68
0.311
11 (6/5)
0.276 (0.286)
0.743 (0.748)
0.892 (0.894)
21.27 (21.98)
0.329 (0.301)
0.319
0.688
0.872
22.00
0.299
Appendix
Table 7: Effect of full interaction in the posed setting. Views are listed as total (context/target), and depth is scored on the context views. FULL encodes only the context views at test time, matching MASK. Gray values in parentheses are from FULL with all views encoded, shown for reference.
Backbone
Method
Camera pose
Depth
NVS
RPA@3 ↑
RPA@5 ↑
RPA@10 ↑
AbsRel ↓
δ<1.25↑
PSNR ↑
LPIPS ↓
–
w/o Pretrained init
36.2
56.6
77.3
0.247
0.652
19.82
0.438
Pixio (base)
Baseline
66.6
81.7
91.7
0.195
0.758
21.48
0.312
w/ Sem. guidance
82.3
90.8
96.0
0.185
0.790
21.55
0.305
w/ Sem. guidance + Proj.
81.6
90.7
96.2
0.189
0.789
21.52
0.305
DINOv3 (base)
Baseline
46.8
64.3
80.0
0.217
0.718
20.23
0.361
Appendix
Table 8: Ablation of semantic guidance on DL3DV dataset. NVS metrics are reported as a reference only, since target views are processed together with the context under full-view attention.
Method
BlendedMVS
ScanNetv2
AbsRel ↓
δ<1.25↑
RPA@5 ↑
RPA@15 ↑
RPA@30 ↑
AbsRel ↓
δ<1.25↑
RPA@5 ↑
RPA@15 ↑
RPA@30 ↑
w/o Pretrained init
0.162
0.800
4.2
16.9
26.7
0.179
0.745
0.7
9.8
28.7
Baseline
0.113
0.891
9.1
25.0
37.7
0.120
0.873
2.9
21.0
47.7
w/ Sem. guidance
0.094
0.919
12.6
33.0
44.3
0.078
0.943
6.2
39.5
67.6
w/ Sem. guidance + Proj.
0.096
0.919
11.8
28.5
41.5
0.082
0.932
6.4
35.7
63.0
Appendix
Table 9: 3D probing with semantic guidance. Self-supervised models from Tab. 8 are frozen, and only the output heads are trained with 3D supervision on each dataset. Each scene uses 10 frames spanning 24 frames on BlendedMVS and 64 frames on ScanNetv2.
Method
EPE ↓
< 1px ↑
< 3px ↑
< 5px ↑
w/o Pretrained init
4.05
25.0
68.0
81.9
Baseline
3.16
32.0
74.7
86.7
w/ Sem. guidance
2.74
31.4
77.4
89.6
w/ Sem. guidance + Proj.
2.66
32.3
78.7
90.2
Appendix
Table 10: Flow probing on StaticThings3D with a frozen backbone, on non-occluded pixels.
Sem. guid.
Asym. align.
Camera pose
Multi-view depth
Single-view depth
RPA@3 ↑
RPA@5 ↑
RPA@10 ↑
AUC@30 ↑
AbsRel ↓
δ<1.25↑
AbsRel ↓
δ<1.25↑
66.6
81.7
91.7
88.0
0.195
0.758
0.230
0.638
✓
69.8
84.6
93.5
89.9
0.191
0.768
0.191
0.729
✓
82.3
90.8
96.0
93.0
0.185
0.790
0.173
0.757
✓
✓
82.8
92.2
96.9
93.7
0.188
0.790
0.165
0.799
Appendix
Table 11: Stacking semantic guidance and asymmetric context alignment on DL3DV.
Figure 15: Camera pose accuracy on DL3DV when adding semantic guidance and asymmetric context alignment to the self-supervised baseline, averaged over 5, 7, 9, and 11 input views.
Method
NVS
Camera pose
Depth
PSNR ↑
LPIPS ↓
RPA@3 ↑
RPA@5 ↑
RPA@10 ↑
AUC@30 ↑
AbsRel ↓
δ<1.25↑
Neural rendering (interpolative)
22.28
0.258
0.0
0.0
1.1
0.26
–
–
Neural rendering (predictive)
22.10
0.258
0.0
0.3
0.8
0.21
–
–
Ours
21.75
0.280
85.7
94.1
98.5
95.12
0.230
0.775
Appendix
Table 12: Comparison with neural rendering methods on DL3DV dataset (8 views spanning 32 frames; 4 context + 4 target). The two neural rendering variants differ in whether target views are sampled between or beyond the context views during training. Our NVS metrics are grayed out, as target views are processed together with the context under full-view attention.
S
Training data
N
RPA@3 ↑
RPA@5 ↑
RPA@10 ↑
RPA@15 ↑
RPA@30 ↑
AUC@30 ↑
64
DL3DV
330
5.5
13.6
33.7
48.1
71.6
42.5
DL3DV + SpatialVID
7.5
17.3
39.2
55.9
76.7
47.9
96
DL3DV
220
5.3
12.5
31.4
46.2
68.4
40.3
DL3DV + SpatialVID
7.3
16.4
36.7
51.9
71.7
44.7
Appendix
Table 13: Zero-shot camera pose accuracy on ORBIT (dynamic) dataset. S is the frame span of each 10-view window, and N is the number of windows.
Figure 16: Camera poses on ORBIT windows spanning 64 frames (left) and 96 frames (right), from models trained on DL3DV only and on DL3DV with SpatialVID. Predictions are shown in color and ground truth in black, aligned by a similarity transform.
Figure 17: Additional qualitative comparison of recovered camera trajectories. In-domain camera poses from 10 unposed views on RE10K (left) and DL3DV (right). Predictions are shown in color and ground truth in black, aligned by a similarity transform.