LucidNFT: LR-Anchored Multi-Reward Preference Optimization for Generative Real-World Super-Resolution

• URL: http://arxiv.org/abs/2603.05947v1
• Date: Fri, 06 Mar 2026 06:30:34 GMT
• Title: LucidNFT: LR-Anchored Multi-Reward Preference Optimization for Generative Real-World Super-Resolution
• Abstract summary: Preference-based reinforcement learning (RL) is a natural fit because each LR input yields a rollout group of candidates to compare.<n>We propose LucidNFT, a multi-reward RL framework for flow-matching Real-ISR.<n>Experiments show that LucidNFT consistently improves strong flow-based Real-ISR baselines.
• Score: 21.290660354883595
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Generative real-world image super-resolution (Real-ISR) can synthesize visually convincing details from severely degraded low-resolution (LR) inputs, yet its stochastic sampling makes a critical failure mode hard to avoid: outputs may look sharp but be unfaithful to the LR evidence (semantic and structural hallucination), while such LR-anchored faithfulness is difficult to assess without HR ground truth. Preference-based reinforcement learning (RL) is a natural fit because each LR input yields a rollout group of candidates to compare. However, effective alignment in Real-ISR is hindered by (i) the lack of a degradation-robust LR-referenced faithfulness signal, and (ii) a rollout-group optimization bottleneck where naive multi-reward scalarization followed by normalization compresses objective-wise contrasts, causing advantage collapse and weakening the reward-weighted updates in DiffusionNFT-style forward fine-tuning. Moreover, (iii) limited coverage of real degradations restricts rollout diversity and preference signal quality. We propose LucidNFT, a multi-reward RL framework for flow-matching Real-ISR. LucidNFT introduces LucidConsistency, a degradation-robust semantic evaluator that makes LR-anchored faithfulness measurable and optimizable; a decoupled advantage normalization strategy that preserves objective-wise contrasts within each LR-conditioned rollout group before fusion, preventing advantage collapse; and LucidLR, a large-scale collection of real-world degraded images to support robust RL fine-tuning. Experiments show that LucidNFT consistently improves strong flow-based Real-ISR baselines, achieving better perceptual-faithfulness trade-offs with stable optimization dynamics across diverse real-world scenarios.

Related papers

• DNF-SR: Dual-Input and Negative-Aware Feature Fine-Tuning for Real-World Image Super-Resolution [50.847827209157494]
  Unmediatedly feeding LR into a diffusion model creates a distributional gap with the model's original input.<n>We propose a Dual-input and Negative-aware Feature fine-tuning method for Real-ISR.<n>We exploit this property and propose a post-training optimization method, Negative-aware Feature Fine-Tuning (NF2T), which guides the model toward producing higher-quality results.
  arXiv  Detail & Related papers  (2026-09-14T06:53:22Z)
• MeanSR: Restoration Trajectory Learning for One-Step Perceptual Super-Resolution [50.60539317126614]
  MeanSR learns an LR-conditioned average velocity field to directly capture the finite-time transition from noisy inputs to plausible HR outputs.<n>Experiments show that MeanSR outperforms CTMSR on CLIPIQA, MUSIQ, and MANIQA while substantially reducing FLOPs and inference latency.
  arXiv  Detail & Related papers  (2026-08-10T10:33:18Z)
• Clipping Bottleneck: Stabilizing RLVR via Stochastic Recovery of Near-Boundary Signals [83.0127582612634]
  Near-boundary Rescue (NSR) is a minimal, plug-and-play modification that retains out-of-bound tokens to recover lost signals.<n>NSR substantially improves training stability and delivers consistent gains over strong baselines such as DAPO and GSPO.
  arXiv  Detail & Related papers  (2026-05-21T16:45:31Z)
• RefReward-SR: LR-Conditioned Reward Modeling for Preference-Aligned Super-Resolution [26.50828843718706]
  RefReward-SR is a low-resolution (LR) reference-aware reward model for preference-aligned SR.<n>It assesses high-resolution (HR) reconstructions conditioned on LR inputs, treating the LR image as a semantic anchor.<n>Our framework achieves substantially better alignment with human judgments, producing reconstructions that preserve semantic consistency.
  arXiv  Detail & Related papers  (2026-03-25T11:20:51Z)
• OARS: Process-Aware Online Alignment for Generative Real-World Image Super-Resolution [47.35253230699505]
  OARS is a process-aware online alignment framework that evaluates the LR to SR transition by jointly modeling fidelity preservation and perceptual gain.<n>We show consistent perceptual improvements while maintaining fidelity, achieving state-of-the-art performance on Real-ISR benchmarks.
  arXiv  Detail & Related papers  (2026-03-13T09:12:46Z)
• Bidirectional Reward-Guided Diffusion for Real-World Image Super-Resolution [79.35296000454694]
  Diffusion-based super-resolution can synthesize rich details, but models trained on synthetic paired data often fail on real-world LR images.<n>We propose Bird-SR, a reward-guided diffusion framework that formulates super-resolution as trajectory-level preference optimization.<n>Experiments on real-world SR benchmarks demonstrate that Bird-SR consistently outperforms state-of-the-art methods in perceptual quality.
  arXiv  Detail & Related papers  (2026-02-05T19:21:45Z)
• Robust-R1: Degradation-Aware Reasoning for Robust Visual Understanding [54.05243949024302]
  Existing robust MLLMs rely on implicit training/adaptation that focuses solely on visual encoder generalization.<n>We propose Robust-R1, a novel framework that explicitly models visual degradations through structured reasoning chains.<n>Our approach integrates: (i) supervised fine-tuning for degradation-aware reasoning foundations, (ii) reward-driven alignment for accurately perceiving degradation parameters, and (iii) dynamic reasoning depth scaling adapted to degradation intensity.
  arXiv  Detail & Related papers  (2025-12-19T12:56:17Z)
• ControlSR: Taming Diffusion Models for Consistent Real-World Image Super Resolution [68.72454974431749]
  We present ControlSR, a new method that can tame Diffusion Models for consistent real-world image super-resolution (Real-ISR)<n>Our model can achieve better performance across multiple metrics on several test sets and generate more consistent SR results with LR images than existing methods.
  arXiv  Detail & Related papers  (2024-10-18T08:35:57Z)
• Co-learning Single-Step Diffusion Upsampler and Downsampler with Two Discriminators and Distillation [28.174638880324014]
  Super-resolution (SR) aims to reconstruct high-resolution (HR) images from their low-resolution (LR) counterparts.<n>We propose a co-learning framework that jointly optimize a single-step diffusion-based upsampler and a learnable downsampler.
  arXiv  Detail & Related papers  (2024-10-10T07:12:46Z)
• ICF-SRSR: Invertible scale-Conditional Function for Self-Supervised Real-world Single Image Super-Resolution [60.90817228730133]
  Single image super-resolution (SISR) is a challenging problem that aims to up-sample a given low-resolution (LR) image to a high-resolution (HR) counterpart. Recent approaches are trained on simulated LR images degraded by simplified down-sampling operators. We propose a novel Invertible scale-Conditional Function (ICF) which can scale an input image and then restore the original input with different scale conditions.
  arXiv  Detail & Related papers  (2023-07-24T12:42:45Z)
• Frequency Consistent Adaptation for Real World Super Resolution [64.91914552787668]
  We propose a novel Frequency Consistent Adaptation (FCA) that ensures the frequency domain consistency when applying Super-Resolution (SR) methods to the real scene. We estimate degradation kernels from unsupervised images and generate the corresponding Low-Resolution (LR) images. Based on the domain-consistent LR-HR pairs, we train easy-implemented Convolutional Neural Network (CNN) SR models.
  arXiv  Detail & Related papers  (2020-12-18T08:25:39Z)

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.