Public

3DGS — Public Tier

(5 scenes)

Full-access development tier with all data visible.

What you get

Measurements (y), ideal forward operator (H), spec ranges, ground truth (x_true), and true mismatch spec.

How to use

Load HDF5 → compare reconstruction vs x_true → check consistency → iterate.

What to submit

Reconstructed signals (x_hat) and corrected spec as HDF5.

Parameter Specifications

✓

True spec visible — use these exact values for Scenario III oracle reconstruction.

Parameter Spec Range True Value Unit
camera_pose -1.0 – 2.0 0.5 mm/deg
focal_length -5.0 – 10.0 2.5 pixels
point_cloud_init -2.0 – 4.0 1.0 mm

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

16.22 dB

SSIM 0.4885

Scenario II (Mismatch)

15.42 dB

SSIM 0.3644

Scenario III (Oracle)

14.86 dB

SSIM 0.4147

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 13.55 0.7184 13.40 0.6942 11.26 0.4810
scene_01 11.41 0.1961 13.08 0.1153 13.99 0.3354
scene_02 19.57 0.6233 17.91 0.5565 16.23 0.4134
scene_03 20.36 0.4162 17.31 0.0917 17.94 0.4288

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 GaussianShader + gradient 0.807 33.42 0.956 0.95 ✓ Certified Wang et al., ICCV 2024
2 2DGS + gradient 0.801 33.3 0.955 0.93 ✓ Certified Huang et al., CVPR 2024
3 3D-GS++ + gradient 0.797 32.88 0.951 0.94 ✓ Certified Kerbl et al., SIGGRAPH 2024
4 NeRFactor2 + gradient 0.793 32.99 0.952 0.91 ✓ Certified Barron et al., NeurIPS 2024
5 3D-GS + gradient 0.761 31.15 0.933 0.88 ✓ Certified Kerbl et al., SIGGRAPH 2023
6 NeRF + gradient 0.758 31.09 0.932 0.87 ✓ Certified Mildenhall et al., ECCV 2020
7 Mesh-GS + gradient 0.731 28.69 0.895 0.93 ✓ Certified Li et al., ECCV 2024
8 Instant-NGP + gradient 0.723 28.84 0.897 0.88 ✓ Certified Muller et al., SIGGRAPH 2022
9 Mip-NeRF 360 + gradient 0.689 26.87 0.855 0.89 ✓ Certified Barron et al., CVPR 2022
10 Photogrammetry + gradient 0.664 25.31 0.812 0.93 ✓ Certified Structure-from-Motion baseline
11 COLMAP+MVS + gradient 0.663 25.35 0.813 0.92 ✓ Certified Schonberger & Frahm, CVPR 2016

Visible Data Fields

y H_ideal spec_ranges x_true true_spec

Dataset

Format: HDF5
Scenes: 5

Scoring Formula

0.4 × PSNR_norm + 0.4 × SSIM + 0.2 × (1 − ‖y − Ĥx̂‖/‖y‖)

PSNR: 40% SSIM: 40% Consistency: 20%
Back to 3DGS