AI Paper Atlas

Research map / Image generation / Image restoration

Unified Removal of Raindrops and Reflections: A New Benchmark and A Novel Pipeline

Xingyu Liu, Zewei He, Yu Chen, Chunyu Zhu, Zixuan Chen, Xing Luo, Zheming Lu

ECCV 2026 · Image generation › Image restoration · approach: Efficient restoration architectures · 0 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
add Modulate&Gate module instead of channel concatenation of I_lq and I_s (ablation)DiffUR3 · RDRF-testingPSNR: 28.60 -> 29.41Improved
add Modulate&Gate module instead of channel concatenation of I_lq and I_s (ablation)DiffUR3 · RDRF-testingLPIPS: 0.0899 -> 0.0813Improved
add Modulate block (ablation)DiffUR3 without Modulate · RDRF-testingPSNR: 29.11 -> 29.41Improved
add Gate block (ablation)DiffUR3 without Gate · RDRF-testingPSNR: 29.12 -> 29.41Improved
use both I_lq and I_s as conditions rather than I_lq only (ablation)DiffUR3 conditioning · RDRF-testingPSNR: 28.33 -> 29.41Improved
use both I_lq and I_s as conditions rather than I_s only (ablation)DiffUR3 conditioning · RDRF-testingPSNR: 28.54 -> 29.41Improved
add Fidelity Encoder to VAE decoding (ablation)DiffUR3 · RDRF-testingPSNR: 26.95 -> 29.41Improved
add Fidelity Encoder to VAE decoding (ablation)DiffUR3 · RDRF-testingSSIM: 0.8042 -> 0.9372Improved
add Fidelity Encoder to VAE decoding (ablation)DiffUR3 · RDRF-testingLPIPS: 0.1082 -> 0.0813Improved
replace CFW (w=1.0) with Fidelity Encoder (ablation)DiffUR3 · RDRF-testingPSNR: 29.18 -> 29.41Improved
replace CFW (w=0.5) with Fidelity Encoder (ablation)DiffUR3 · RDRF-testingPSNR: 28.78 -> 29.41Improved
add stage II diffusion generation on top of stage I restorationDRSformer stage I · RDRF-testingLPIPS: 0.0925 -> 0.0813Improved

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