atlas

Research map / Image generation / Image restoration

Semi-Supervised Parametric Real-World Image Harmonization

CVPR 2023 · Image generation › Image restoration · approach: Image harmonization and color transformation · 30 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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
Use dual-stream training with unpaired realistic composites in addition to artist-retouched supervised data (ablation)Parametric image harmonization model · RealHM 512×512MSE: 264.0 -> 154.3Helped
Add the local shading map to global RGB curves (ablation)Dual-stream harmonization model · RealHM 512×512MSE: 222.6 -> 154.3Helped
Use dual-stream training with realistic composites (ablation)Global-curves-plus-shading model · RealHM 512×512PSNR: 26.89 -> 28.40Helped
Add the local shading map (ablation)Dual-stream harmonization model · RealHM 512×512B-T score: 0.217 -> 0.252Helped
Replace synthetic-only supervised training with dual-stream training on artist-retouched pairs and unpaired realistic compositesParametric image harmonization model · Real composites user studyB-T score: 0.2257 -> 0.3025Helped
Use the proposed model rather than HarmonizerImage harmonization · iHarmony 2048×2048 Adobe5kPSNR: 38.33 -> 38.29No clear effect

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