arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.02436cs.CV
收藏

几何感知隐式记忆用于视频世界模型

Geometry-Aware Implicit Memory for Video World Models

School of Intelligence Science and Technology, Nanjing University(南京大学智能科学与技术学院);Kling Team, Kuaishou Technology(快手技术 Kling 团队);Tsinghua University(清华大学)

Zhengxuan Wei, Xu Guo, Xinghui Li, Xunzhi Xiang, Min Wei, Yiran Zhu, Qiulin Wang, Xintao Wang, Pengfei Wan, Xiangwang Hou, Qi Fan

AI总结:

提出GIM-World框架,通过轻量级Transformer编码器将可变长度历史压缩为固定大小的记忆令牌,并利用相机可查询的几何头在训练期间从冻结的基础模型中蒸馏3D场景结构,从而在长时程视频生成中保持几何和视觉一致性。

摘要:

视频世界模型旨在模拟可控的视觉环境,但长时程展开取决于模型在观察离开其原生上下文窗口后记住的内容。显式记忆保留帧或在线3D重建,可能会遭受启发式检索错误、冗余外观存储或重建伪影。隐式记忆将历史压缩为紧凑状态,但现有设计没有明确约束以编码跨视图场景几何。我们提出GIM-World,一种用于视频世界模型的几何感知隐式记忆框架。轻量级Transformer编码器将可变长度历史压缩为固定大小的记忆令牌,相机可查询的几何头在训练期间从冻结的基础模型中将3D场景结构蒸馏到记忆中,信息引导的剪枝规则在历史增长时保持编码成本有界。在推理时丢弃几何教师,留下轻量级记忆模块。在MIND上的实验表明,GIM-World在保持长时程几何和视觉一致性方面优于显式和隐式记忆基线。

英文摘要:

Video world models aim to simulate controllable visual environments, but long-horizon rollouts depend on what the model remembers after observations leave its native context window. Explicit memories retain frames or online 3D reconstructions, which can suffer from heuristic retrieval errors, redundant appearance storage, or reconstruction artifacts. Implicit memories compress history into a compact state, but existing designs are not explicitly constrained to encode cross-view scene geometry. We propose GIM-World, a geometry-aware implicit memory framework for video world models. A lightweight transformer encoder compresses variable-length history into fixed-size memory tokens, a camera-queryable geometry head distills 3D scene structure from a frozen foundation model into the memory during training, and an information-guided pruning rule keeps encoding cost bounded as history grows. The geometry teacher is discarded at inference, leaving a lightweight memory module. Experiments on MIND show that GIM-World better preserves long-horizon geometric and visual consistency than both explicit- and implicit-memory baselines.

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。

↑