Allo{SR}$^2$: Rectifying One-Step Super-Resolution to Stay Real via Allomorphic Generative Flows

2026-04-21

AI Summary

This work addresses trajectory deviation, artifacts, and generative prior collapse in one-step image super-resolution caused by the absence of multi-step refinement. To this end, we propose a single-step super-resolution method based on heterogeneous generative flows, integrating three key mechanisms: SNR-guided trajectory initialization, Flow-Anchored Trajectory Consistency (FATC), and Heterogeneous Trajectory Matching (ATM). These components jointly enable high-fidelity restoration and high-realism generation within a single inference step. Our approach is the first to unify flow-based generative priors, velocity-level supervision, and self-adversarial distribution alignment within a coherent vector field formulation. Extensive experiments demonstrate state-of-the-art performance on both synthetic and real-world benchmarks for one-step super-resolution, achieving an optimal trade-off among efficiency, fidelity, and generative quality.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Real-world image super-resolution (Real-SR) has been revolutionized by leveraging the powerful generative priors of large-scale diffusion and flow-based models. However, fine-tuning these models on limited LR-HR pairs often precipitates "prior collapse" that the model sacrifices its inherent generative richness to overfit specific training degradations. This issue is further exacerbated in one-step generation, where the absence of multi-step refinement leads to significant trajectory drift and artifact generation. In this paper, we propose Allo{SR}$^2$, a novel framework that rectifies one-step SR trajectories via allomorphic generative flows to maintain high-fidelity generative realism. Specifically, we utilize Signal-to-Noise Ratio (SNR) Guided Trajectory Initialization to establish a physically grounded starting state by aligning the degradation level of LR latent features with the optimal anchoring timestep of the pre-trained flow. To ensure a stable, curvature-free path for one-step inference, we propose Flow-Anchored Trajectory Consistency (FATC), which enforces velocity-level supervision across intermediate states. Furthermore, we develop Allomorphic Trajectory Matching (ATM), a self-adversarial alignment strategy that minimizes the distributional discrepancy between the SR flow and the generative flow in a unified vector field. Extensive experiments on both synthetic and real-world benchmarks demonstrate that Allo{SR}$^2$ achieves state-of-the-art performance in one-step Real-SR, offering a superior balance between restoration fidelity and generative realism while maintaining extreme efficiency.

Problem

• Real-world Super-Resolution
• One-step Generation
• Prior Collapse
• Trajectory Drift
• Generative Realism

Innovation

• one-step super-resolution
• generative flows
• trajectory rectification
• prior collapse
• distributional alignment