# Geo-Align：基于度量几何奖励的视频生成对齐

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-05-22 08:00
- AIHOT 分数：48
- AIHOT 链接：https://aihot.virxact.com/items/cmpkp4pvu08xysl01ed20lzxm
- 原文链接：https://arxiv.org/abs/2605.23903

## AI 摘要

针对现有摄像机控制视频重渲染方法因缺乏真实多视角数据而泛化能力有限的问题，Geo-Align 提出了首个专门用于此任务的强化学习框架。该框架基于预训练模型，通过尺度感知的感知奖励进行优化。其核心是引入度量3D估计器从生成视频中提取精确摄像机轨迹，并显式惩罚旋转与平移的偏差。同时，设计了基于真实条件视频和合成目标轨迹的数据管道策略，摆脱了对配对数据的依赖。实验表明，Geo-Align 在精确摄像机控制与视觉保真度上均优于现有的监督学习基线。

## 正文

Camera-controlled video generation has achieved remarkable progress in recent years. However, existing video-to-video re-rendering methods primarily rely on Supervised Fine-Tuning using synthetic datasets. At present, there is an extreme scarcity of synchronized, multi-view real-world video data. Consequently, the prevailing paradigm often exhibits limited generalization when processing out-of-distribution real-world videos, with models struggling to accurately adhere to physical scales and camera trajectories. To bridge this gap, we propose Geo-Align, the first Reinforcement Learning framework specifically designed for camera-controlled video re-rendering. Built upon a pretrained model, we optimize the model through a scale-aware perceptual reward mechanism. Specifically, we introduce a metric 3D estimator to extract precise camera trajectories from generated videos, explicitly penalizing deviations in rotation and translation. Furthermore, we meticulously designed a data pipeline strategy based on real-world conditioning videos and target camera trajectories derived from synthetic data, eliminating the reliance on paired data. Extensive experiments demonstrate that Geo-Align consistently outperforms existing supervised learning baselines in both precise camera controllability and visual fidelity, indicating the effectiveness of our method.