Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2602.08068 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908820587937792 |
|---|---|
| author | Li, Chunyang Yang, Yuanbo Shao, Jiahao Zhou, Hongyu Schwarz, Katja Liao, Yiyi |
| author_facet | Li, Chunyang Yang, Yuanbo Shao, Jiahao Zhou, Hongyu Schwarz, Katja Liao, Yiyi |
| contents | Video generation with controllable camera viewpoints is essential for applications such as interactive content creation, gaming, and simulation. Existing methods typically adapt pre-trained video models using camera poses relative to a fixed reference, e.g., the first frame. However, these encodings lack shift-invariance, often leading to poor generalization and accumulated drift. While relative camera pose embeddings defined between arbitrary view pairs offer a more robust alternative, integrating them into pre-trained video diffusion models without prohibitive training costs or architectural changes remains challenging. We introduce ReRoPE, a plug-and-play framework that incorporates relative camera information into pre-trained video diffusion models without compromising their generation capability. Our approach is based on the insight that Rotary Positional Embeddings (RoPE) in existing models underutilize their full spectral bandwidth, particularly in the low-frequency components. By seamlessly injecting relative camera pose information into these underutilized bands, ReRoPE achieves precise control while preserving strong pre-trained generative priors. We evaluate our method on both image-to-video (I2V) and video-to-video (V2V) tasks in terms of camera control accuracy and visual fidelity. Our results demonstrate that ReRoPE offers a training-efficient path toward controllable, high-fidelity video generation. See project page for more results: https://sisyphe-lee.github.io/ReRoPE/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08068 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ReRoPE: Repurposing RoPE for Relative Camera Control Li, Chunyang Yang, Yuanbo Shao, Jiahao Zhou, Hongyu Schwarz, Katja Liao, Yiyi Computer Vision and Pattern Recognition Video generation with controllable camera viewpoints is essential for applications such as interactive content creation, gaming, and simulation. Existing methods typically adapt pre-trained video models using camera poses relative to a fixed reference, e.g., the first frame. However, these encodings lack shift-invariance, often leading to poor generalization and accumulated drift. While relative camera pose embeddings defined between arbitrary view pairs offer a more robust alternative, integrating them into pre-trained video diffusion models without prohibitive training costs or architectural changes remains challenging. We introduce ReRoPE, a plug-and-play framework that incorporates relative camera information into pre-trained video diffusion models without compromising their generation capability. Our approach is based on the insight that Rotary Positional Embeddings (RoPE) in existing models underutilize their full spectral bandwidth, particularly in the low-frequency components. By seamlessly injecting relative camera pose information into these underutilized bands, ReRoPE achieves precise control while preserving strong pre-trained generative priors. We evaluate our method on both image-to-video (I2V) and video-to-video (V2V) tasks in terms of camera control accuracy and visual fidelity. Our results demonstrate that ReRoPE offers a training-efficient path toward controllable, high-fidelity video generation. See project page for more results: https://sisyphe-lee.github.io/ReRoPE/ |
| title | ReRoPE: Repurposing RoPE for Relative Camera Control |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.08068 |