AI Paper Atlas

Research map / Image generation / Super-resolution

Fast Bayesian Uncertainty Estimation and Reduction of Batch Normalized Single Image Super-Resolution Network

Aupendu Kar, Prabir Kumar Biswas

CVPR 2021 · Image generation › Super-resolution · approach: CNN and transformer SR architectures · 12 citations (OpenAlex)

Check 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
Generate MC samples in a single forward pass rather than repeated MCBN inferenceBayesian VDSR · Set14 face, 5–15 MC samplesrelative inference time: 14.28 vs 1.0; 33.48 vs 1.97; 52.67 vs 2.96Improved
Apply uncertainty reduction (ablation)BayesianVDSR · BSD100 adversarial attacks, levels 1–16, ×2/×3/×4PSNR: 31.16/ 28.36/ 27.03 -> 31.55/ 28.63/ 27.16; 18.04/ 18.37/ 18.82 -> 24.78/ 24.64/ 23.94Improved
Apply uncertainty reduction (ablation)BayesianVDSR · BSD100, no attack, ×2/×3/×4PSNR: 32.02/ 28.95/ 27.43 -> 31.94/ 28.87/ 27.32Hurt
Apply uncertainty reduction (ablation)BayesianVDSR · Set5, Gaussian noise level 5, ×2/×3/×4PSNR: 33.44/ 31.28/ 29.70 -> 34.08/ 31.69/ 29.78Mixed
Apply uncertainty reduction (ablation)BayesianVDSR · Set5, Gaussian noise levels 10 and 15, ×2/×3/×4PSNR: 29.32/ 28.06/ 27.18 -> 31.68/ 29.73/ 28.17; 26.43/ 25.61/ 25.04 -> 29.65/ 28.31/ 27.12Improved
Apply uncertainty reduction (ablation)BayesianVDSR · Noisy LR images with Poisson or Speckle noise, ×2/×3/×4PSNR: 29.16/ 28.04/ 27.10 -> 31.28/ 29.49/ 28.10; 27.80/ 26.83/ 26.22 -> 29.73/ 28.58/ 27.45Improved
Increase the number of MC samplesBayesian VDSR · BSD100PSNRImproved

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