AI Paper Atlas

Research map / Image generation / Super-resolution

Breaking Scale Anchoring: Frequency Representation Learning for Accurate High-Resolution Inference from Low-Resolution Training

Wenshuo Wang, Fan Zhang

ICLR 2026 · Image generation › Super-resolution · approach: Perceptual and real-world SR · 1 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 Frequency Representation Learning (multi-resolution training, Nyquist-normalized frequency encoding, frequency consistency loss)seven architectures (GNN, Transformer, CNN, Diffusion, NO, Neural ODE, NN) · 3D HIT fluid simulation, 129^3RMSE: 3.57x-7.74x lower RMSE; RMSE ratio 1.017-1.338 -> 0.135-0.375Improved
add FRL to WeatherGFT TransformerWeatherGFT · ERA5 Z500 7-day, 1440x721ACC: 0.44 -> 0.65Improved
add FRL to ClimODEClimODE · ERA5 Z500 7-day, 1440x721RMSE: 772 -> 502Improved
increase test resolution without FRLbaseline STF models · 3D HIT fluid simulationRMSE ratio: 1.017-1.338 (error stays anchored)No clear effect
increase test resolution beyond training Nyquist (ablation)baseline models on 2D convection-diffusion · Error Ratiobandlimited/wideband error ratio: 1.000 (64x64) -> 0.223 (GNN, 256x256)Hurt
FRL training cost overheadFRL variants · 3D fluid simulationtraining time / memory: 1.1x-1.4x time, 1.3x-1.5x memory, <2% inference overheadHurt

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