通过自监督反馈提升图像超分辨率，深层细节更真实
OracleZoom: On-Policy Self-Distillation Inspired Reference-Constrained Recursive Image Super Resolution
- 用自身预测结果递归优化，结合参考图像约束
- 跨尺度监督+无参考质量目标，深层放大效果提升0.713 CLIPIQA
- 适合需要高倍放大的图像修复与生成任务
递归超分辨率（SR）通过反复将模型输出回传实现极端放大，类似逐级缩放。但深层放大时真实标签难以获取，导致深层预测缺乏监督。本文提出OracleZoom，一种基于策略自蒸馏、参考约束的递归框架，在训练过程中保留最后一帧真实信息以突破监督边界。直接与跨尺度监督约束可验证内容，无参考质量目标引导未解析的细粒度细节。使用KL约束的预训练潜在先验限制质量漂移，EMA一致性稳定监督边界。在七个数据集上，OracleZoom在各缩放尺度均达当前最优，平均CLIPIQA为0.713，深层放大增益更显著，且大幅减少幻觉。代码、数据与模型已公开。
原文摘要 · Abstract (English)
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/ .
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。