arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.24240cs.CV
收藏

联合几何与轨迹一致性学习用于一步现实世界超分辨率

Joint Geometric and Trajectory Consistency Learning for One-Step Real-World Super-Resolution

University of Electronic Science（电子科学大学）；Tsinghua University（清华大学）；Nanjing University（南京大学）

Chengyan Deng, Zhangquan Chen, Li Yu, Kai Zhang, Xue Zhou, Wang Zhang

阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

GTASR通过引入轨迹对齐和双参考结构校正，解决现实世界超分辨率中的几何解耦和一致性漂移问题，实现高效且高质量的一步生成。

摘要：

基于扩散的现实世界图像超分辨率（Real-ISR）在感知质量上取得了显著成就，但因其迭代采样而面临高计算成本的问题。尽管最近的蒸馏方法利用大规模文本到图像（T2I）先验使一步生成成为可能，但通常受限于参数数量庞大以及教师模型固有的能力限制。作为轻量级替代方案，一致性模型提供高效的推理，但面临两个关键限制：传递训练中固有的一致性漂移累积现象，以及我们称之为“几何解耦”的现象——生成轨迹在像素层面实现对齐却无法保持结构一致性。为了解决这些挑战，我们提出了GTASR（几何轨迹对齐超分辨率），一种针对Real-ISR的简单而有效的一致性训练范式。具体来说，我们引入了轨迹对齐（TA）策略以通过全路径投影校正切向量场，并引入双参考结构校正（DRSR）机制以施加严格结构约束。广泛实验验证了GTASR在保持最小延迟的同时优于代表性基线的性能。代码和模型将在https://github.com/Blazedengcy/GTASR上发布。

英文摘要：

Diffusion-based Real-World Image Super-Resolution (Real-ISR) achieves impressive perceptual quality but suffers from high computational costs due to iterative sampling. While recent distillation approaches leveraging large-scale Text-to-Image (T2I) priors have enabled one-step generation, they are typically hindered by prohibitive parameter counts and the inherent capability bounds imposed by teacher models. As a lightweight alternative, Consistency Models offer efficient inference but struggle with two critical limitations: the accumulation of consistency drift inherent to transitive training, and a phenomenon we term "Geometric Decoupling" - where the generative trajectory achieves pixel-wise alignment yet fails to preserve structural coherence. To address these challenges, we propose GTASR (Geometric Trajectory Alignment Super-Resolution), a simple yet effective consistency training paradigm for Real-ISR. Specifically, we introduce a Trajectory Alignment (TA) strategy to rectify the tangent vector field via full-path projection, and a Dual-Reference Structural Rectification (DRSR) mechanism to enforce strict structural constraints. Extensive experiments verify that GTASR delivers superior performance over representative baselines while maintaining minimal latency. The code and model will be released at https://github.com/Blazedengcy/GTASR.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

↑