AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

RestoreVAR: Visual Autoregressive Generation for All-in-One Image Restoration

Sudarshan Rajagopalan, Kartik Narayan, Vishal M. Patel

ICLR 2026 · Image generation › Autoregressive and latent generation · approach: High-resolution diffusion synthesis

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 latent refiner transformer (LRT) to convert discrete to continuous latents (ablation)RestoreVAR · RESIDE testPSNR: 21.71 -> 24.67Improved
use LRT instead of HART diffusion refiner (ablation)RestoreVAR · RESIDE testPSNR: 23.48 -> 24.67Improved
feed final transformer block output z to LRT via cross-attention (ablation)LRT · RESIDE testPSNR: 21.23 -> 24.67Improved
add LRT refiner (ablation)RestoreVAR · perceptual quality (MUSIQ/CLIPIQA)MUSIQ / CLIPIQA: 66.56/0.472 -> 66.45/0.481No clear effect
condition on continuous latent of degraded image instead of discrete multi-scale latents (ablation)RestoreVAR · validation accuracyvalidation accuracyImproved
add PatchGAN discriminator to VAE decoder fine-tuning (ablation)VAE decoder · reconstruction qualitysharpness (qualitative)Improved
replace absolute positional embeddings with 2D RoPE (ablation)RestoreVAR transformer · 512x512 fine-tuningvalidation accuracyImproved
fine-tune VAE decoder on continuous latents onlyVAR VQVAE decoder · reconstruction (1000 samples)PSNR (dB): 22.59 -> 28.14Improved
use VAR-based RestoreVAR instead of LDM-based AiOR methodsgenerative AiOR · RESIDEPSNR: 24.48 -> 24.67Improved
use RestoreVAR instead of LDM-based methods for inference speedAiOR · inference time per imagetime (s): 8.477 -> 0.201Improved
compare RestoreVAR to LDM-based methods on unseen-degradation perceptual metricsRestoreVAR · unseen/mixed degradations averageMUSIQ: 56.767 (AutoDir) vs 56.217No clear effect
compare RestoreVAR to non-generative methods on pixel fidelityRestoreVAR · RESIDEPSNR: 32.02 (PromptIR) vs 24.67Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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