L3 Benchmark ⊙ Testnet L3-004-T1

CACTI

Coded Aperture Compressive Temporal Imaging

🚀 Submit a solution — mine PWM View benchmark MD →
Lineage: L1 Principle · L1-004 → L2 Spec · L2-004 → L3 CACTI Benchmark

How the CACTI forward model works

A coded aperture shutters the scene at sub-frame rates within one exposure; the detector integrates the shutter-modulated frames into a single 2-D measurement. Recovering the high-speed video frames from one coded snapshot is the inverse problem.

video
f(x,y,t)
→
shutter C_t
(binary)
→
Σ over t
(detector)
→
2-D snapshot
y(x,y)

Standard benchmark

nominal · no mismatch

One benchmark, evaluated on the full dataset — no separate public / hidden / dev splits. CACTI-6 standard 256x256x8 kobe/traffic/drop/runner/crash/aerial; calibrated mask (no mismatch).

26.0

Success threshold ε

256×256

Resolution

3

Reference baselines

Reference baselines shipped with the benchmark

Solver Quality Q Status
baseline:EfficientSCI 0.920 ⊙ testnet
baseline:PnP-FFDNet 0.780 ⊙ testnet
baseline:GAP-TV 0.750 ⊙ testnet

Community leaderboard 0 submitted

No community solutions yet.

🚀 Be the first — submit a solver →
⚠

Mismatch is its own benchmark + spec

The standard benchmark above assumes a perfectly-calibrated system (all mismatch parameters = 0). Real sensors have calibration error — mask shift, rotation, dispersion-slope drift. Those are tested under a separate mismatch spec and benchmark, where Ω includes the mismatch dimensions (mask_dx, mask_dy, mask_theta, disp_a1_error, disp_alpha_error).

Mismatch Spec · L2-004 → Mismatch benchmark tiers →