AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

FVAR: Next-Focus Prediction for Visual Autoregressive Modeling

Xiaofan Li, Chenming Wu, Yanpeng Sun, Jiaming Zhou, Delin Qu, Yansong Qu, Weihao Bo, Haibao Yu, Dingkang Liang

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images

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 uniform downsampling pyramid with progressive refocusing (defocus PSF) (ablation)FVAR-d16 · ImageNet 256x256FID: 3.51 -> 2.89Improved
replace uniform downsampling with progressive refocusing (ablation)FVAR-d16 · ImageNet 1024x1024FID: 8.15 -> 6.85Improved
use physics-based defocus PSF instead of Gaussian blur (ablation)FVAR-d16 · ImageNet 256x256FID: 3.32 -> 2.89Improved
Gaussian blur instead of standard downsampling (ablation)VAR-d16 · ImageNet 1024x1024FID: 8.15 -> 7.50Improved
add high-frequency residual teacher with distillation (ablation)FVAR-d16 · ImageNet 256x256FID: 3.06 -> 2.89Improved
add high-frequency residual teacher (ablation)FVAR-d16 · ImageNet 1024x1024FID: 7.20 -> 6.85Improved
use separate dual tokenizers for structure and alias (ablation)FVAR-d16 · ImageNet 256x256FID: 3.14 -> 2.89Improved
use separate dual tokenizers (ablation)FVAR-d16 · ImageNet 1024x1024FID: 7.40 -> 6.85Improved
FVAR on VAR-d12 versus VAR-d12VAR-d12 · ImageNet 256x256FID: 5.81 -> 3.95Improved
FVAR versus M-VAR at d24 (no added parameters)VAR-d24 · ImageNet 256x256FID: 1.93 -> 1.75Improved
FVAR at 512x512 versus VARVAR-d36 · ImageNet 512x512FID: 2.63 -> 2.28Improved

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