Public

Compressed Ultrafast Photography (CUP) — 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
dmd_encoding_error -0.4 – 0.8 0.2 -
streak_sweep_calibration -1.0 – 2.0 0.5 -
temporal_spatial_coupling -2.0 – 4.0 1.0 -

InverseNet Baseline Scores

Method: CPU_baseline — Mismatch parameter: nominal

Scenario I (Ideal)

7.84 dB

SSIM 0.3015

Scenario II (Mismatch)

7.84 dB

SSIM 0.1991

Scenario III (Oracle)

10.94 dB

SSIM 0.3682

Per-scene breakdown (4 scenes)
Scene PSNR I SSIM I PSNR II SSIM II PSNR III SSIM III
scene_00 7.64 0.2894 7.66 0.2019 10.79 0.3739
scene_01 7.65 0.3040 7.65 0.1960 10.77 0.3632
scene_02 8.07 0.3073 8.08 0.2002 11.17 0.3690
scene_03 8.02 0.3052 7.95 0.1985 11.04 0.3666

Public Tier Leaderboard

# Method Score PSNR SSIM Consistency Trust Source
1 DiffusionCUP + gradient 0.868 38.8 0.985 0.93 ✓ Certified Qiao et al., Nat. Photonics 2020 (updated 2024)
2 STFormer-CUP + gradient 0.841 36.15 0.974 0.95 ✓ Certified Wang et al., CVPR 2022 (CUP)
3 DAUHST-CUP + gradient 0.828 35.91 0.973 0.9 ✓ Certified Cai et al., NeurIPS 2022 (CUP)
4 PnP-FastDVDnet + gradient 0.787 32.57 0.949 0.91 ✓ Certified Tassano et al., CVPR 2020 (CUP)
5 E2E-CNN-CUP + gradient 0.767 31.72 0.94 0.87 ✓ Certified Liang et al., CVPR 2019
6 DeSCI-CUP + gradient 0.751 30.02 0.917 0.92 ✓ Certified Liu et al., IEEE TPAMI 2018 (CUP adapt.)
7 GAP-TV + gradient 0.673 26.2 0.838 0.88 ✓ Certified Yuan, ICSIP 2016
8 TwIST-CUP + gradient 0.632 24.19 0.775 0.9 ✓ Certified Bioucas-Dias & Figueiredo, IEEE TIP 2007 (CUP)
9 TV-CUP + gradient 0.577 22.26 0.701 0.87 ✓ Certified Gao et al., Nature 2014

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 Compressed Ultrafast Photography (CUP)