arXiv cs.AI· Yukuo Ma, Cong Liu, Junke Wang, Junqi Liu, Haibin Huang, Zuxuan Wu, Xuelong Li·· 12 小时前AI 评分39
ReCam：用全上下文记忆实现稳定长视频生成
Recast the Past: Stable Long Video Generation with Full Context Memory
AI 导读
ReCam 将历史潜变量的局部时空块分组为向量并直接投影到 Diffusion Transformer 的高维隐藏空间，通过原生自注意力使用记忆 token，在降低 token 数量的同时保留对完整历史的块级访问。
正文 · 原文
Abstract:Long-horizon video generation requires retaining an ever-expanding history, but conditioning on its growing token sequence is costly. Existing methods restrict or compress historical context, potentially losing information needed for future generation. This work proposes ReCam, which groups local spatiotemporal blocks of historical latents into vectors and projects them directly into the high-dimensional hidden space of a Diffusion Transformer. The model learns to use the resulting memory tokens through native self-attention, reducing token count while retaining block-level access to the full history. To mitigate temporal drift during autoregressive rollout, we further introduce dual-pointer denoising with separate noise schedules for recent context and target frames. Among the evaluated methods, ReCam achieves the highest VBench-Long overall score of 80.1 and competitive static-scene revisitation performance on R2M-Bench. It also leads on all four consistency metrics for off-screen subject reappearance, including clothing consistency of 4.09 versus 3.03 for the best baseline.
| Comments: | Code: this https URL |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.27986 [cs.CV] |
| (or arXiv:2610.27986v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27986 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yukuo Ma [view email]
[v1]
Tue, 20 Oct 2026 10:27:07 UTC (28,973 KB)
来源：arXiv cs.AI · arxiv.org