LPNSR: Prior-Enhanced Diffusion Image Super-Resolution via LR-Guided Noise Prediction

2026-03-21

AI Summary

Diffusion models for image super-resolution often struggle to balance efficiency and perceptual quality due to unconstrained Gaussian noise and bicubic upsampling initialization. This work proposes LPNSR, a novel framework operating under a residual-shifted diffusion paradigm, which derives—for the first time—a closed-form optimal solution for intermediate noise. To incorporate structural priors, we introduce a low-resolution (LR)-guided multi-input perceptual noise predictor. Additionally, a high-quality pre-upsampling network is designed and trained end-to-end to refine the initial state. The resulting method achieves state-of-the-art perceptual quality on both synthetic and real-world datasets with only four denoising steps, without relying on large-scale text-to-image priors.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Diffusion-based image super-resolution (SR), which aims to reconstruct high-resolution (HR) images from corresponding low-resolution (LR) observations, faces a fundamental trade-off between inference efficiency and reconstruction quality. The state-of-the-art residual-shifting diffusion framework achieves efficient 4-step inference, yet suffers from severe performance degradation in compact sampling trajectories. This is mainly attributed to two core limitations: the inherent suboptimality of unconstrained random Gaussian noise in intermediate steps, which leads to error accumulation and insufficient LR prior guidance, and the initialization bias caused by naive bicubic upsampling. In this paper, we propose LPNSR, a prior-enhanced efficient diffusion framework to address these issues. We first mathematically derive the closed-form analytical solution of the optimal intermediate noise for the residual-shifting diffusion paradigm, and accordingly design an LR-guided multi-input-aware noise predictor to replace random Gaussian noise, embedding LR structural priors into the reverse process while fully preserving the framework's core efficient residual-shifting mechanism. We further mitigate initial bias with a high-quality pre-upsampling network to optimize the diffusion starting point. With a compact 4-step trajectory, LPNSR can be optimized in an end-to-end manner. Extensive experiments demonstrate that LPNSR achieves state-of-the-art perceptual performance on both synthetic and real-world datasets, without relying on any large-scale text-to-image priors. The source code of our method can be found at https://github.com/Faze-Hsw/LPNSR.

Problem

• image super-resolution
• diffusion model
• inference efficiency
• reconstruction quality
• LR prior

Innovation

• diffusion-based super-resolution
• noise prediction
• LR-guided prior
• residual-shifting
• efficient sampling