AI Paper Atlas

Research map / Image generation / Super-resolution

IFCSR: Inference-Free Fidelity-Realism Control for One-Step Diffusion-based Real-World Image Super-Resolution

Jonghee Back, Jongju Kim, Jeong-Uk Kim, Eunjin Kim, Minyong Jeon

CVPR 2026 · Image generation › Super-resolution · approach: Diffusion-based super-resolution

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
increase blending parameter gamma from 0 to 1 (image-space blend of fidelity and realism outputs) (ablation)IFCSR · RealSR (x4)PSNR: 28.00 -> 22.10Hurt
increase blending parameter gamma from 0 to 1 (ablation)IFCSR · RealSR (x4)CLIPIQA: 0.3337 -> 0.7011Improved
increase blending parameter gamma from 0 to 1 (ablation)IFCSR · DRealSR (x4)PSNR: 30.86 -> 24.63Hurt
increase blending parameter gamma from 0 to 1 (ablation)IFCSR · RealSR (x4)LPIPS: 0.2921 -> 0.2540 -> 0.3501Mixed
use two-stage training instead of single-stage training with the baseline LPIPS+MSE loss (ablation)IFCSR · RealSR (x4)fidelity-realism spectrum alignmentImproved
replace baseline LPIPS+MSE loss with depth-weighted specialized losses in two-stage training (ablation)IFCSR · RealSR (x4)fidelity-realism spectrum widthImproved
use depth-dependent weighting in the feature loss (ablation)IFCSR · RealSR (x4)LPIPS monotonicity across gammaHurt
blend fidelity and realism outputs in image space instead of latent-space controlIFCSR · control costextra network inferences per controlImproved
blend two outputs with differing structure in image spaceIFCSR · failure casevisual artifactsHurt
generate both fidelity- and realism-specific images with two LoRA networksIFCSR · Efficiency, RTX 3090 FP16inference time (s): 0.20 (PiSA-SR) -> 0.35Hurt
use IFCSR at gamma = 1.0 instead of PiSA-SR on DRealSRcontrollable one-step ISR · DRealSR (x4)CLIPIQA: 0.6975 -> 0.7446Improved

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