AI Paper Atlas

Research map / Image generation / Super-resolution

Toward Bridging the Simulated-to-Real Gap: Benchmarking Super-Resolution on Real Data

Thomas Köhler, Michel Bätz, Farzad Naderi, André Kaup, Andreas Maier, Christian Rieß

TPAMI 2019 · Image generation › Super-resolution · approach: Efficient lightweight 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
Evaluate on real hardware-binned LR data instead of simulated bicubic LR data19 SR algorithms (SISR and MFSR) · SupER global motion, 2x2 and 4x4 binningSpearman rank correlation between simulated and real rankings: rho < 0.6Hurt
Use deep networks (VDSR, DRCN) instead of classical SR methods on real dataSISR · SupER real dataPSNR/SSIM/IFC rankNo clear effect
Use multi-frame instead of single-image SR under pure global motionSR algorithms · SupER global motionnormalized PSNR and IFCImproved
Use multi-frame instead of single-image SR under mixed motionSR algorithms · SupER mixed motionnormalized PSNR and IFCHurt
Use multi-frame instead of single-image SR under local motionSR algorithms · SupER local motionnormalized PSNR and IFCHurt
Use external training data instead of self-exemplarsSISR · SupER global motionnormalized PSNR and IFCImproved
Add H.265/HEVC compression to LR input (QP up to 40)all SR algorithms · SupER global motion, 3xnormalized PSNR and IFCHurt
Add photometric outlier frames to the input setMFSR · SupER global motion, 3x, K=11 framesnormalized PSNR and IFCHurt
Use reconstruction-based robust models (IRWSR, BEPSR) instead of interpolation under photometric variationsMFSR · SupER photometric variationnormalized PSNR and IFCImproved
Increase magnification factor in observer studyObserver agreement · SupER observer studyKendall coefficient of agreement uImproved
Omit sanity checks (tolerate all 8 errors, n_f = 8) (ablation)Observer study · SupER observer studyinter-observer variance: n_f = 8 vs n_f = 1Hurt
Use LPIPS or IFC instead of PSNR/SSIM/MS-SSIM as quality measureFull-reference measures · SupER vs B-T observer scoresweighted Kendall tau distanceImproved

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