AI Paper Atlas

Research map / Image generation / Image restoration

Zero-Order Reverse Filtering

Xin Tao, Chao Zhou, Xiaoyong Shen, Jue Wang, Jiaya Jia

ICCV 2017 · Image generation › Image restoration · approach: Diffusion-based universal restoration · 34 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
apply zero-order reverse filtering iterations to Gaussian-filtered imagesfiltered image · BSD300-based, GSGT PSNR (dB): 27.75 -> 41.70Improved
apply zero-order reverse filtering to bilateral-filtered imagesfiltered image · BSD300-based, BFGT PSNR (dB): 25.50 -> 45.28Improved
apply zero-order reverse filtering to guided-filtered imagesfiltered image · BSD300-based, GFGT PSNR (dB): 28.21 -> 51.05Improved
apply zero-order reverse filtering to median-filtered imagesfiltered image · BSD300-based, MFGT PSNR (dB): 26.01 -> 26.07No clear effect
apply zero-order reverse filtering to weighted median filtered imagesfiltered image · BSD300-based, WMFGT PSNR (dB): 33.00 -> 22.98 (final), 35.43 (best)Mixed
apply zero-order reverse filtering to local extrema filtered imagesfiltered image · BSD300-based, LEGT PSNR (dB): 29.30 -> 7.96 (final), 44.21 (best)Mixed
run 50 iterations instead of stopping at best iteration for partially reversible bilateral grid (ablation)zero-order reverse filtering · BSD300-based, BFGGT PSNR (dB): 37.70 -> 35.78Hurt
run 50 iterations instead of stopping at best iteration for domain transform filter (ablation)zero-order reverse filtering · BSD300-based, RFGT PSNR (dB): 31.16 -> 28.63Hurt
run 50 iterations instead of stopping at best iteration for permutohedral lattice (ablation)zero-order reverse filtering · BSD300-based, BFPLGT PSNR (dB): 36.92 -> 32.97Hurt
add reverse filtering to initialization with bicubic upsamplingbicubic x2 super-resolution · single example imagePSNR (dB): 28.99 -> 30.50Improved
use zero-order reverse filtering instead of Lanczos3 interpolationx2 super-resolution · single example imagePSNR (dB): 29.57 -> 30.50Improved
add reverse filtering as post-processing to ScSR and SRCNN outputslearning-based super-resolution · single example imagePSNRImproved

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