AI Paper Atlas

Research map / 3D vision / Depth and stereo

Learned Compressive Representations for Single-Photon 3D Imaging

Felipe Gutierrez-Barragan, Fangzhou Mu, Andrei Ardelean, Atul Ingle, Claudio Bruschini, Edoardo Charbon, Yin Li, Mohit Gupta, Andreas Velten

ICCV 2023 · 3D vision › Depth and stereo · approach: Spike and event camera reconstruction · 11 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
learn the coding tensors end to end instead of using fixed Fourier codes (ablation)compressive histogram layer · simulated Middlebury test setmean absolute depth errorImproved
use random unoptimized coding tensors instead of learned ones (ablation)compressive histogram layer · simulated scene SBR=0.1, 64x and 128x compressiondepth reconstruction qualityHurt
use spatio-temporal rather than temporal-only coding tensors (ablation)learned compressive histograms · low SBR, compression ratio above 100depth errorImproved
increase spatial block size of coding tensors at 64x compression (ablation)learned separable coding tensors · low vs high SBRMAE and detail preservationMixed
make coding tensor separable (1024x4x4) (ablation)learned coding tensors · 128x compressiondepth performance: 2,097,152 -> 133,120 paramsNo clear effect
further reduce parameters of coding tensors (fewer than 10,000) (ablation)learned separable coding tensors · 128x compressiondepth reconstruction sharpnessHurt
use learned compressive histograms instead of peak compression oracle (in-pixel argmax depth)compressive histograms · low SBR, 64x compressiondepth reconstructionImproved
quantize coding tensors C to 8 bits (ablation)compressive histograms · UltraPhase implementationdepth performanceNo clear effect
use compressive histograms instead of coarse histograms on-chipUltraPhase processing · power analysiscompute power share of totalMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy