IRPO: Boosting Image Restoration via Post-training GRPO

• URL: http://arxiv.org/abs/2512.00814v2
• Date: Tue, 09 Dec 2025 06:21:17 GMT
• Title: IRPO: Boosting Image Restoration via Post-training GRPO
• Authors: Haoxuan Xu, Yi Liu, Boyuan Jiang, Jinlong Peng, Donghao Luo, Xiaobin Hu, Shuicheng Yan, Haoang Li,
• Abstract summary: We propose IRPO, a low-level GRPO-based post-training paradigm.<n>We first explore a data formulation principle for low-level post-training paradigm.<n>We then model a reward-level criteria system that balances objective accuracy and human perceptual preference.
• Score: 59.588079259093035
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent advances in post-training paradigms have achieved remarkable success in high-level generation tasks, yet their potential for low-level vision remains rarely explored. Existing image restoration (IR) methods rely on pixel-level hard-fitting to ground-truth images, struggling with over-smoothing and poor generalization. To address these limitations, we propose IRPO, a low-level GRPO-based post-training paradigm that systematically explores both data formulation and reward modeling. We first explore a data formulation principle for low-level post-training paradigm, in which selecting underperforming samples from the pre-training stage yields optimal performance and improved efficiency. Furthermore, we model a reward-level criteria system that balances objective accuracy and human perceptual preference through three complementary components: a General Reward for structural fidelity, an Expert Reward leveraging Qwen-VL for perceptual alignment, and a Restoration Reward for task-specific low-level quality. Comprehensive experiments on six in-domain and five out-of-domain (OOD) low-level benchmarks demonstrate that IRPO achieves state-of-the-art results across diverse degradation types, surpassing the AdaIR baseline by 0.83 dB on in-domain tasks and 3.43 dB on OOD settings. Our code can be shown in https://github.com/HaoxuanXU1024/IRPO.

Related papers

• Simplicity Prevails: The Emergence of Generalizable AIGI Detection in Visual Foundation Models [15.709482146201283]
  A simple linear classifier trained on the frozen features of modern Vision Foundation Models establishes a new state-of-the-art.<n>We show that this baseline matches specialized detectors on standard benchmarks but also decisively outperforms them on in-the-wild datasets.<n>We conclude by advocating for a paradigm shift in AI forensics, moving from overfitting on static benchmarks to harnessing the evolving world knowledge of foundation models for real-world reliability.
  arXiv  Detail & Related papers  (2026-02-02T07:20:02Z)
• Implicit Neural Representation-Based Continuous Single Image Super Resolution: An Empirical Study [50.15623093332659]
  Implicit neural representation (INR) has become the standard approach for arbitrary-scale image super-resolution (ASSR)<n>We compare existing techniques across diverse settings and present aggregated performance results on multiple image quality metrics.<n>We examine a new loss function that penalizes intensity variations while preserving edges, textures, and finer details during training.
  arXiv  Detail & Related papers  (2026-01-25T07:09:20Z)
• RS-Prune: Training-Free Data Pruning at High Ratios for Efficient Remote Sensing Diffusion Foundation Models [14.093802378976315]
  Diffusion-based remote sensing (RS) generative foundation models rely on large amounts of globally representative data.<n>We propose a training-free, two-stage data pruning approach that quickly select a high-quality subset under high pruning ratios.<n> Experiments show that, even after pruning 85% of the training data, our method significantly improves convergence and generation quality.
  arXiv  Detail & Related papers  (2025-12-29T06:44:06Z)
• OmniQuality-R: Advancing Reward Models Through All-Encompassing Quality Assessment [55.59322229889159]
  We propose OmniQuality-R, a unified reward modeling framework that transforms multi-task quality reasoning into continuous and interpretable reward signals.<n>We use a reasoning-enhanced reward modeling dataset to form a reliable chain-of-thought dataset for supervised fine-tuning.<n>We evaluate OmniQuality-R on three key IQA tasks: aesthetic quality assessment, technical quality evaluation, and text-image alignment.
  arXiv  Detail & Related papers  (2025-10-12T13:46:28Z)
• LRPO: Enhancing Blind Face Restoration through Online Reinforcement Learning [54.51101908523586]
  Blind Face Restoration (BFR) encounters inherent challenges in exploring its large solution space.<n>We propose a Likelihood-Regularized Policy Optimization (LRPO) framework, the first to apply online reinforcement learning (RL) to the BFR task.<n>Our proposed LRPO significantly improves the face restoration quality over baseline methods and achieves state-of-the-art performance.
  arXiv  Detail & Related papers  (2025-09-27T14:42:29Z)
• BranchGRPO: Stable and Efficient GRPO with Structured Branching in Diffusion Models [57.304411396229035]
  We present BranchGRPO, a method that restructures the rollout process into a branching tree.<n>On HPDv2.1 image alignment, BranchGRPO improves alignment scores by up to textbf16% over DanceGRPO.<n>A hybrid variant, BranchGRPO-Mix, further accelerates training to 4.7x faster than DanceGRPO without degrading alignment.
  arXiv  Detail & Related papers  (2025-09-07T12:53:06Z)
• NatureGAIA: Pushing the Frontiers of GUI Agents with a Challenging Benchmark and High-Quality Trajectory Dataset [16.676904484703]
  We introduce NaturalGAIA, a novel benchmark engineered on the principle of Causal Pathways.<n>This paradigm structures complex tasks into a series of verifiable atomic steps, ensuring rigorous, fully automated, and reproducible standard for assessment.<n>We then utilize this dataset to perform Reinforcement FineTuning (RFT) on the Q2.5-VL-7B model.
  arXiv  Detail & Related papers  (2025-08-02T11:53:41Z)
• IQPFR: An Image Quality Prior for Blind Face Restoration and Beyond [56.99331967165238]
  Blind Face Restoration (BFR) addresses the challenge of reconstructing degraded low-quality (LQ) facial images into high-quality (HQ) outputs.<n>We propose a novel framework that incorporates an Image Quality Prior (IQP) derived from No-Reference Image Quality Assessment (NR-IQA) models.<n>Our method outperforms state-of-the-art techniques across multiple benchmarks.
  arXiv  Detail & Related papers  (2025-03-12T11:39:51Z)
• Pseudo-triplet Guided Few-shot Composed Image Retrieval [20.040511832864503]
  Composed Image Retrieval (CIR) is a challenging task that aims to retrieve the target image with a multimodal query. We propose a novel two-stage pseudo triplet guided few-shot CIR scheme, dubbed PTG-FSCIR. In the first stage, we propose an attentive masking and captioning-based pseudo triplet generation method, to construct pseudo triplets from pure image data. In the second stage, we propose a challenging triplet-based CIR fine-tuning method, where we design a pseudo modification text-based sample challenging score estimation strategy.
  arXiv  Detail & Related papers  (2024-07-08T14:53:07Z)
• Hyperspectral Benchmark: Bridging the Gap between HSI Applications through Comprehensive Dataset and Pretraining [11.935879491267634]
  Hyperspectral Imaging (HSI) serves as a non-destructive spatial spectroscopy technique with a multitude of potential applications. A recurring challenge lies in the limited size of the target datasets, impeding exhaustive architecture search. This study introduces an innovative benchmark dataset encompassing three markedly distinct HSI applications.
  arXiv  Detail & Related papers  (2023-09-20T08:08:34Z)
• A Practical Contrastive Learning Framework for Single-Image Super-Resolution [51.422185656787285]
  We investigate contrastive learning-based single image super-resolution from two perspectives. We propose a practical contrastive learning framework for SISR, named PCL-SR. Compared with existing benchmark methods, we re-train them by our proposed PCL-SR framework and achieve superior performance.
  arXiv  Detail & Related papers  (2021-11-27T15:42:12Z)

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.