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

• URL: http://arxiv.org/abs/2512.09282v1
• Date: Wed, 10 Dec 2025 03:10:52 GMT
• Title: FoundIR-v2: Optimizing Pre-Training Data Mixtures for Image Restoration Foundation Model
• Authors: Xiang Chen, Jinshan Pan, Jiangxin Dong, Jian Yang, Jinhui Tang,
• Abstract summary: We propose a high-capacity diffusion-based image restoration foundation model, FoundIR-v2.<n>By leveraging the data mixing law, our method ensures a balanced dataset composition.<n>Our method achieves favorable performance against state-of-the-art approaches.
• Score: 89.23522479092537
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• ToReMi: Topic-Aware Data Reweighting for Dynamic Pre-Training Data Selection [28.75333303894706]
  ToReMi is a novel framework that adjusts training sample weights according to their topical associations and observed learning patterns.<n>Our experiments reveal that ToReMi variants consistently achieve superior performance over conventional pre-training approaches.
  arXiv  Detail & Related papers  (2025-04-01T12:06:42Z)
• Dynamic Loss-Based Sample Reweighting for Improved Large Language Model Pretraining [55.262510814326035]
  Existing reweighting strategies primarily focus on group-level data importance.<n>We introduce novel algorithms for dynamic, instance-level data reweighting.<n>Our framework allows us to devise reweighting strategies deprioritizing redundant or uninformative data.
  arXiv  Detail & Related papers  (2025-02-10T17:57:15Z)
• FoundIR: Unleashing Million-scale Training Data to Advance Foundation Models for Image Restoration [66.61201445650323]
  Existing methods suffer from a generalization bottleneck in real-world scenarios.<n>We contribute a million-scale dataset with two notable advantages over existing training data.<n>We propose a robust model, FoundIR, to better address a broader range of restoration tasks in real-world scenarios.
  arXiv  Detail & Related papers  (2024-12-02T12:08:40Z)
• Semi-Supervised Fine-Tuning of Vision Foundation Models with Content-Style Decomposition [4.192370959537781]
  We present a semi-supervised fine-tuning approach designed to improve the performance of pre-trained foundation models on downstream tasks with limited labeled data. We evaluate our approach on multiple datasets, including MNIST, its augmented variations, CIFAR-10, SVHN, and GalaxyMNIST.
  arXiv  Detail & Related papers  (2024-10-02T22:36:12Z)
• Minimizing Embedding Distortion for Robust Out-of-Distribution Performance [1.0923877073891446]
  We introduce a novel approach we call "similarity loss", which can be incorporated into the fine-tuning process of any task. We evaluate our approach on two diverse tasks: image classification on satellite imagery and face recognition.
  arXiv  Detail & Related papers  (2024-09-11T19:22:52Z)
• What Matters When Repurposing Diffusion Models for General Dense Perception Tasks? [49.84679952948808]
  Recent works show promising results by simply fine-tuning T2I diffusion models for dense perception tasks.<n>We conduct a thorough investigation into critical factors that affect transfer efficiency and performance when using diffusion priors.<n>Our work culminates in the development of GenPercept, an effective deterministic one-step fine-tuning paradigm tailed for dense visual perception tasks.
  arXiv  Detail & Related papers  (2024-03-10T04:23:24Z)
• Learning Energy-Based Models by Cooperative Diffusion Recovery Likelihood [64.95663299945171]
  Training energy-based models (EBMs) on high-dimensional data can be both challenging and time-consuming. There exists a noticeable gap in sample quality between EBMs and other generative frameworks like GANs and diffusion models. We propose cooperative diffusion recovery likelihood (CDRL), an effective approach to tractably learn and sample from a series of EBMs.
  arXiv  Detail & Related papers  (2023-09-10T22:05:24Z)
• Ensemble Modeling for Multimodal Visual Action Recognition [50.38638300332429]
  We propose an ensemble modeling approach for multimodal action recognition. We independently train individual modality models using a variant of focal loss tailored to handle the long-tailed distribution of the MECCANO [21] dataset.
  arXiv  Detail & Related papers  (2023-08-10T08:43:20Z)
• An Empirical Study on Distribution Shift Robustness From the Perspective of Pre-Training and Data Augmentation [91.62129090006745]
  This paper studies the distribution shift problem from the perspective of pre-training and data augmentation. We provide the first comprehensive empirical study focusing on pre-training and data augmentation.
  arXiv  Detail & Related papers  (2022-05-25T13:04:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.