# FoundIR-v2: Optimizing Pre-Training Data Mixtures for Image Restoration Foundation Model

arXiv: https://arxiv.org/abs/2512.09282 · HTML: https://mlsanity.com/paper/2512.09282/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Placed in Inverse problems and image restoration, one of 14 directions in Diffusion models.
- Builds on 52 of its 64 references that are in the map.
- First posted on arXiv 10 Dec 2025.

## Placed in

- Diffusion models / [Inverse problems and image restoration](https://mlsanity.com/diffusion/inverse-problems-and-image-restoration/)
- Mixture-of-experts / [Vision mixture-of-experts](https://mlsanity.com/moe/vision-mixture-of-experts/)

## Builds on (most cited first, 30 of 52)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://mlsanity.com/paper/1812.04948/) (2018)
- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [SGDR: Stochastic Gradient Descent with Warm Restarts](https://mlsanity.com/paper/1608.03983/) (2016)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Elucidating the Design Space of Diffusion-Based Generative Models](https://mlsanity.com/paper/2206.00364/) (2022)
- [Restormer: Efficient Transformer for High-Resolution Image Restoration](https://mlsanity.com/paper/2111.09881/) (2021)
- [Enhanced Deep Residual Networks for Single Image Super-Resolution](https://mlsanity.com/paper/1707.02921/) (2017)
- [Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data](https://mlsanity.com/paper/2401.10891/) (2024)
- [MUSIQ: Multi-scale Image Quality Transformer](https://mlsanity.com/paper/2108.05997/) (2021)
- [Exploring CLIP for Assessing the Look and Feel of Images](https://mlsanity.com/paper/2207.12396/) (2022)
- [Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data](https://mlsanity.com/paper/2107.10833/) (2021)
- [Pre-Trained Image Processing Transformer](https://mlsanity.com/paper/2012.00364/) (2020)
- [Designing a Practical Degradation Model for Deep Blind Image Super-Resolution](https://mlsanity.com/paper/2103.14006/) (2021)
- [MANIQA: Multi-dimension Attention Network for No-Reference Image Quality Assessment](https://mlsanity.com/paper/2204.08958/) (2022)
- [Exploiting Diffusion Prior for Real-World Image Super-Resolution](https://mlsanity.com/paper/2305.07015/) (2023)
- [SeeSR: Towards Semantics-Aware Real-World Image Super-Resolution](https://mlsanity.com/paper/2311.16518/) (2023)
- [Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild](https://mlsanity.com/paper/2401.13627/) (2024)
- [Pixel-Aware Stable Diffusion for Realistic Image Super-resolution and Personalized Stylization](https://mlsanity.com/paper/2308.14469/) (2023)
- [One-Step Effective Diffusion Network for Real-World Image Super-Resolution](https://mlsanity.com/paper/2406.08177/) (2024)
- [TransWeather: Transformer-based Restoration of Images Degraded by Adverse Weather Conditions](https://mlsanity.com/paper/2111.14813/) (2021)
- [From Patches to Pictures (PaQ-2-PiQ): Mapping the Perceptual Space of Picture Quality](https://mlsanity.com/paper/1912.10088/) (2019)
- [Learning A Sparse Transformer Network for Effective Image Deraining](https://mlsanity.com/paper/2303.11950/) (2023)
- [Image Restoration with Mean-Reverting Stochastic Differential Equations](https://mlsanity.com/paper/2301.11699/) (2023)
- [InstructIR: High-Quality Image Restoration Following Human Instructions](https://mlsanity.com/paper/2401.16468/) (2024)
- [NH-HAZE: An Image Dehazing Benchmark with Non-Homogeneous Hazy and Haze-Free Images](https://mlsanity.com/paper/2005.03560/) (2020)
- [Dense Haze: A benchmark for image dehazing with dense-haze and haze-free images](https://mlsanity.com/paper/1904.02904/) (2019)
- [Depicting Beyond Scores: Advancing Image Quality Assessment through Multi-modal Language Models](https://mlsanity.com/paper/2312.08962/) (2023)

## Built on by (most cited first, 9 of 9)

- [TIR-Agent: Training an Explorative and Efficient Agent for Image Restoration](https://mlsanity.com/paper/2603.27742/) (2026)
- [LoViF 2026 Challenge on Real-World All-in-One Image Restoration: Methods and Results](https://mlsanity.com/paper/2604.19445/) (2026)
- [TPGDiff: Hierarchical Triple-Prior Guided Diffusion for Image Restoration](https://mlsanity.com/paper/2601.20306/) (2026)
- [PixRestore: Unified Image Restoration via Pixel Diffusion Transformer](https://mlsanity.com/paper/2608.16793/) (2026)
- [Position: Evaluation of Visual Processing Should Be Human-Centered, Not Metric-Centered](https://mlsanity.com/paper/2603.00643/) (2026)
- [The Second LoViF 2026 Challenge on Real-World All-in-One Image Restoration: Methods and Results](https://mlsanity.com/paper/2607.21118/) (2026)
- [DVANet: Degradation-aware Visual-prior Alignment Network for Image Restoration](https://mlsanity.com/paper/2606.19097/) (2026)
- [Learning Polarization Image Restoration with General Restoration Priors](https://mlsanity.com/paper/2609.30728/) (2026)
- [LensBridge: Frequency-Guided Compound Degradation Adaptation for Lens Aberration Correction and Veiling Glare Removal](https://mlsanity.com/paper/2610.00318/) (2026)

## Abstract (arXiv)

Recent studies have witnessed significant advances in image restoration foundation models driven by improvements in the scale and quality of pre-training data. In this work, we find that the data mixture proportions from different restoration tasks are also a critical factor directly determining the overall performance of all-in-one image restoration models. To this end, we propose a high-capacity diffusion-based image restoration foundation model, FoundIR-v2, which adopts a data equilibrium scheduling paradigm to dynamically optimize the proportions of mixed training datasets from different tasks. By leveraging the data mixing law, our method ensures a balanced dataset composition, enabling the model to achieve consistent generalization and comprehensive performance across diverse tasks. Furthermore, we introduce an effective Mixture-of-Experts (MoE)-driven scheduler into generative pre-training to flexibly allocate task-adaptive diffusion priors for each restoration task, accounting for the distinct degradation forms and levels exhibited by different tasks. Extensive experiments demonstrate that our method can address over 50 sub-tasks across a broader scope of real-world scenarios and achieves favorable performance against state-of-the-art approaches.