AI Paper Atlas

Research map / Image generation / Super-resolution

SROBB: Targeted Perceptual Loss for Single Image Super-Resolution

Mohammad Saeed Rad, Behzad Bozorgtabar, Urs-Viktor Marti, Max Basler, Hazim Kemal Ekenel, Jean-Philippe Thiran

ICCV 2019 · Image generation › Super-resolution · approach: CNN and transformer SR architectures · 176 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
replace standard perceptual loss with targeted perceptual loss (ablation)SRGAN-style decoder with MSE and adversarial loss · Ablation user study, 35 imagesuser votes: 417 -> 1212Improved
add targeted perceptual loss on top of MSE-only training (ablation)SR decoder (MSE + adversarial) · Ablation user study, 35 imagesuser votes: 49 -> 1212Improved
targeted perceptual loss wins images by majority vote versus others (ablation)SR decoder · Ablation user study, 35 imagesimages won by majority: 30 of 35 vs 5Improved
use pixel-wise loss only on sky-dominated images (ablation)SR decoder (MSE only) · Ablation user study, sky imagesuser votesMixed
use SROBB instead of ESRGANSR methods · User study, 35 COCO-Stuff imagesuser votes: 436 -> 617Improved
use SROBB instead of SFT-GAN (without segmentation map at test time)SR methods · User study, 35 COCO-Stuff imagesuser votes: 223 -> 617Improved
use SROBB instead of SRGANSR methods · User study, 35 COCO-Stuff imagesuser votes: 201 -> 617Improved
use SROBB instead of RCANSR methods · User study, 35 COCO-Stuff imagesuser votes: 33 -> 617Improved
use SROBB instead of SRGANSR methods · Set5 babyLPIPS: 0.112 -> 0.104Improved
use SROBB instead of SRGANSR methods · Set14 baboonLPIPS: 0.220 -> 0.245Hurt
use SROBB instead of SRGANSR methods · Set5 babyPSNR: 28.413 -> 28.869Improved
use SROBB instead of LapSRNSR methods · Set5 baby / Set14 baboonPSNR: 32.019 -> 28.869Hurt

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