AI Paper Atlas

Research map / Image generation / Super-resolution

WEVSR: Video Diffusion Generators for Real-World Video Super‑Resolution with Wavelet-Enhanced VAE Encoder

Yuying Chen, Zhirui Liu, Linyan Jiang, Qifan Gao, Xianguo Zhang, Jianhou Gan, Wenqi Ren

ICML 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
add timestep sampling (ablation)baseline · Ablation (synthetic eval)PSNR: 22.54 -> 23.82Improved
add noise augmentation (ablation)baseline · Ablation (synthetic eval)PSNR: 22.54 -> 24.65Improved
add wavelet encoder (ablation)baseline · Ablation (synthetic eval)PSNR: 22.54 -> 25.43Improved
add wavelet encoder (ablation)baseline · Ablation (synthetic eval)LPIPS: 0.233 -> 0.168Improved
add timestep sampling on top of noise and wavelet (ablation)noise + wavelet · Ablation (synthetic eval)PSNR: 27.57 -> 28.10Improved
add timestep sampling on top of noise and wavelet (ablation)noise + wavelet · Ablation (synthetic eval)LPIPS: 0.143 -> 0.130Improved
add timestep sampling on top of noise and wavelet (ablation)noise + wavelet · Ablation (synthetic eval)SSIM: 0.839 -> 0.845Improved
add wavelet front-end to WAN VAE encoderWAN VAE · Video VAE reconstructionPSNR: 33.40 -> 33.65Improved
add wavelet front-end to WAN VAE encoderWAN VAE · Video VAE reconstructionrFVD: 3.38 -> 2.96Improved
WEVSR versus SeedVR2-7BWEVSR · UDM10LPIPS: 0.203 -> 0.130Improved
WEVSR versus SeedVR2-7BWEVSR · UDM10PSNR: 26.26 -> 28.10Improved
WEVSR versus SeedVR2-7B on VideoLQWEVSR · VideoLQMUSIQ: 45.76 -> 54.13Mixed

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