# Navigating Image Restoration with VAR's Distribution Alignment Prior

arXiv: https://arxiv.org/abs/2412.21063 · HTML: https://mlsanity.com/paper/2412.21063/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Dec 2024.

## Built on by (most cited first, 10 of 10)

- [Structural and Statistical Texture Knowledge Distillation and Learning for Segmentation](https://mlsanity.com/paper/2503.08043/) (2025)
- [DSwinIR: Rethinking Window-based Attention for Image Restoration](https://mlsanity.com/paper/2504.04869/) (2025)
- [Fast and Accurate Image Restoration and Generation with Rank Enhanced Linear Attention](https://mlsanity.com/paper/2505.16157/) (2025)
- [ScaleWeaver: Weaving Efficient Controllable T2I Generation with Multi-Scale Reference Attention](https://mlsanity.com/paper/2510.14882/) (2025)
- [HSI-VAR: Rethinking Hyperspectral Restoration through Spatial-Spectral Visual Autoregression](https://mlsanity.com/paper/2602.00749/) (2026)
- [Generative Latent Kernel Modeling for Blind Motion Deblurring](https://mlsanity.com/paper/2507.09285/) (2025)
- [StepVAR: Structure-Texture Guided Pruning for Visual Autoregressive Models](https://mlsanity.com/paper/2603.01757/) (2026)
- [Hierarchical Image Tokenization for Multi-Scale Image Super Resolution](https://mlsanity.com/paper/2605.14891/) (2026)
- [Next-Acceleration-Scale Prediction for Autoregressive MRI Reconstruction](https://mlsanity.com/paper/2605.19354/) (2026)
- [Detail Continuation over a Trustworthy Coarse Scale for Autoregressive Super-Resolution](https://mlsanity.com/paper/2608.01823/) (2026)

## Abstract (arXiv)

Generative models trained on extensive high-quality datasets effectively capture the structural and statistical properties of clean images, rendering them powerful priors for transforming degraded features into clean ones in image restoration. VAR, a novel image generative paradigm, surpasses diffusion models in generation quality by applying a next-scale prediction approach. It progressively captures both global structures and fine-grained details through the autoregressive process, consistent with the multi-scale restoration principle widely acknowledged in the restoration community. Furthermore, we observe that during the image reconstruction process utilizing VAR, scale predictions automatically modulate the input, facilitating the alignment of representations at subsequent scales with the distribution of clean images. To harness VAR's adaptive distribution alignment capability in image restoration tasks, we formulate the multi-scale latent representations within VAR as the restoration prior, thus advancing our delicately designed VarFormer framework. The strategic application of these priors enables our VarFormer to achieve remarkable generalization on unseen tasks while also reducing training computational costs. Extensive experiments underscores that our VarFormer outperforms existing multi-task image restoration methods across various restoration tasks.