AI Paper Atlas

Research map / Image generation / Image restoration

Low-Rank Residual Diffusion Models

Junfu Tan, Jiang Yuan

CVPR 2026 · Image generation › Image restoration · approach: Diffusion priors for denoising and deblurring

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
add low-rank residual constraint (ablation)RDDM · Rain1400PSNR: 32.21 -> 33.30Improved
add low-rank residual constraint (ablation)RDDM · GoProSSIM: 0.963 -> 0.927Hurt
add low-rank residual constraint (ablation)RDDM · GoProPSNR: 32.40 -> 32.48No clear effect
add low-rank residual constraint (ablation)ResShift · ISTDPSNR: 24.92 -> 25.23Improved
add low-rank residual constraint (ablation)ResShift · Rain1400PSNR: 25.41 -> 25.44No clear effect
use polynomial decrease rank schedule instead of polynomial increase (p=3) (ablation)LRDM · Rain1400PSNR: 32.88 -> 33.39Improved
use polynomial decrease schedule instead of linear increase (ablation)LRDM · RaindropPSNR: 32.95 -> 33.03Improved
use fixed medium rank (about 80) instead of low or full rank (ablation)LRDM · RaindropPSNR: 31.7 (rank 10) -> 33.03 (rank 80)Mixed
lower polynomial order p from 3 to 1 in decreasing schedule (ablation)LRDM · Rain1400PSNR: 33.39 -> 34.39Improved
lower polynomial order p from 3 to 1 in decreasing schedule (ablation)LRDM · RaindropPSNR: 33.03 -> 32.57Hurt
use LRDM (PD, 2 sampling steps) with few steps (ablation)LRDM · Rain1400PSNR: 34.65 dB at 2 stepsImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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