发表机构
University of Maryland, Baltimore County（马里兰大学巴尔的摩县分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
OracleZoom提出一种受在策略蒸馏启发的参考约束递归超分辨率框架，通过携带真实标签证据并采用无参考质量目标，在七个数据集上实现最先进的超分辨率质量，平均CLIPIQA达0.713，并显著减少幻觉。
AI 中文摘要
递归超分辨率（SR）通过将预测结果反复反馈到同一模型，将固定尺度超分辨率扩展到极端放大倍数，类似于反复放大图像。然而，在每个尺度上，尤其是在深层尺度上，获取真实标签仍然具有挑战性，因为所需的源分辨率呈几何级数增长，导致深层预测缺乏监督。我们提出了OracleZoom，一个受在策略蒸馏启发的、参考约束的框架，该框架在其轨迹上进行训练，同时将最后的真实标签证据携带到监督边界之外。直接和跨尺度监督约束了可验证的内容，而无参考质量目标则引导未解决的细粒度细节。KL约束的预训练潜在先验限制了质量驱动的漂移，而EMA一致性稳定了监督边界。在七个数据集上，OracleZoom在缩放尺度上实现了最先进的SR质量，平均CLIPIQA为0.713，在更深尺度上增益更大，同时显著减少了幻觉。代码、数据和模型可在该https URL获取。
英文摘要
Recursive Super-Resolution (SR) extends fixed-scale SR to extreme magnification by repeatedly feeding predictions back into the same model, analogous to zooming an image repeatedly. However, ground truth availability at every scale, especially at depth, remains challenging as the required source resolution grows geometrically, leaving deeper predictions unsupervised. We present OracleZoom, an on-policy distillation-inspired, reference-constrained framework that trains on its trajectory while carrying the last ground-truth evidence beyond the supervision boundary. Direct and cross-scale supervision constrain verifiable content, while a no-reference quality objective guides unresolved fine-scale detail. A KL-constrained pretrained latent prior limits quality-driven drift, while EMA consistency stabilizes the supervision boundary. Across seven datasets, OracleZoom achieves the state-of-the-art SR quality across zooming scales, averaging 0.713 CLIPIQA, with larger gains on deeper scales, while significantly reducing hallucinations. Code, data, and models are available at https://dipta007.github.io/OracleZoom/ .
