AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

RL-based Stateful Neural Adaptive Sampling and Denoising for Real-Time Path Tracing

Antoine Scardigli, Lukas Cavigelli, Lorenz K Muller

NeurIPS 2023 · Image generation › Fast diffusion sampling · approach: Distillation for few-step generation

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
train sampling network with RL instead of gradient approximation (ablation)ours · Averaged over 4 scenes, 4 sppPSNR (dB): 28.4 -> 28.7Improved
adaptive sampling instead of uniform sampling (ablation)ours · Averaged over 4 scenes, 4 sppPSNR (dB): 27.9 -> 28.7Improved
add latent space encoder with warped latent temporal feedback (ablation)ours · Averaged over 4 scenes, 4 sppPSNR (dB): 28.0 -> 28.7Improved
add temporal feedback and latent encoder (vs none) (ablation)ours · Averaged over 4 scenes, 4 sppPSNR (dB): 27.4 -> 28.7Improved
keep individual sampled pixel values instead of averaging (ablation)ours · Averaged over 4 scenes, 4 sppPSNR (dB): 28.2 -> 28.7Improved
scale down all three networks (ablation)ours · Averaged over 4 scenes, 4 sppPSNR (dB): 28.7 -> 27.3Hurt
use individual sampled values and latent encoder with gradient-approx sampling vs NTASours-A1 vs NTAS · Averaged over 4 scenes, 4 sppPSNR (dB): 27.7 -> 28.4Improved
learned state encoder vs exponential-average heuristicours-A2 vs IMCD · Averaged over 4 scenes, 4 sppPSNR (dB): 27.9 -> 27.9No clear effect
learned sampling and latent update vs probabilistic heuristicsours vs ReSTIR+OIDN · Averaged over 4 scenes, 4 sppPSNR (dB): 28.0 -> 28.7Improved
use 2 spp budget vs strongest baseline at 4 sppours · Averaged over 4 scenes, 4 sppPSNR and latency: 0.2 dB higher PSNR, 1.6x latency reductionImproved
equal 4 spp budget vs strongest baselineours · Averaged over 4 scenes, 4 sppPSNR (dB): +0.6 dB for 3 ms (5.5%) latency increaseImproved
shrink the denoiser network (ablation)ours · EmeraldPlace, 100 ms budgetPSNR (dB)Hurt

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