Tri-Prompting: Video Diffusion with Unified Control over Scene, Subject, and Motion
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910054881427456 |
|---|---|
| author | Zhou, Zhenghong Zhan, Xiaohang Chen, Zhiqin Kim, Soo Ye Zhao, Nanxuan Zheng, Haitian Liu, Qing Zhang, He Lin, Zhe Zhou, Yuqian Luo, Jiebo |
| author_facet | Zhou, Zhenghong Zhan, Xiaohang Chen, Zhiqin Kim, Soo Ye Zhao, Nanxuan Zheng, Haitian Liu, Qing Zhang, He Lin, Zhe Zhou, Yuqian Luo, Jiebo |
| contents | Recent video diffusion models have made remarkable strides in visual quality, yet precise, fine-grained control remains a key bottleneck that limits practical customizability for content creation. For AI video creators, three forms of control are crucial: (i) scene composition, (ii) multi-view consistent subject customization, and (iii) camera-pose or object-motion adjustment. Existing methods typically handle these dimensions in isolation, with limited support for multi-view subject synthesis and identity preservation under arbitrary pose changes. This lack of a unified architecture makes it difficult to support versatile, jointly controllable video. We introduce Tri-Prompting, a unified framework and two-stage training paradigm that integrates scene composition, multi-view subject consistency, and motion control. Our approach leverages a dual-condition motion module driven by 3D tracking points for background scenes and downsampled RGB cues for foreground subjects. To ensure a balance between controllability and visual realism, we further propose an inference ControlNet scale schedule. Tri-Prompting supports novel workflows, including 3D-aware subject insertion into any scenes and manipulation of existing subjects in an image. Experimental results demonstrate that Tri-Prompting significantly outperforms specialized baselines such as Phantom and DaS in multi-view subject identity, 3D consistency, and motion accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_15614 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Tri-Prompting: Video Diffusion with Unified Control over Scene, Subject, and Motion Zhou, Zhenghong Zhan, Xiaohang Chen, Zhiqin Kim, Soo Ye Zhao, Nanxuan Zheng, Haitian Liu, Qing Zhang, He Lin, Zhe Zhou, Yuqian Luo, Jiebo Computer Vision and Pattern Recognition Recent video diffusion models have made remarkable strides in visual quality, yet precise, fine-grained control remains a key bottleneck that limits practical customizability for content creation. For AI video creators, three forms of control are crucial: (i) scene composition, (ii) multi-view consistent subject customization, and (iii) camera-pose or object-motion adjustment. Existing methods typically handle these dimensions in isolation, with limited support for multi-view subject synthesis and identity preservation under arbitrary pose changes. This lack of a unified architecture makes it difficult to support versatile, jointly controllable video. We introduce Tri-Prompting, a unified framework and two-stage training paradigm that integrates scene composition, multi-view subject consistency, and motion control. Our approach leverages a dual-condition motion module driven by 3D tracking points for background scenes and downsampled RGB cues for foreground subjects. To ensure a balance between controllability and visual realism, we further propose an inference ControlNet scale schedule. Tri-Prompting supports novel workflows, including 3D-aware subject insertion into any scenes and manipulation of existing subjects in an image. Experimental results demonstrate that Tri-Prompting significantly outperforms specialized baselines such as Phantom and DaS in multi-view subject identity, 3D consistency, and motion accuracy. |
| title | Tri-Prompting: Video Diffusion with Unified Control over Scene, Subject, and Motion |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.15614 |