Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.16937 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911690972463104 |
|---|---|
| author | Zuo, Yi Wu, Huimin Li, Lingling Liu, Fang Jiao, Licheng Li, Qing |
| author_facet | Zuo, Yi Wu, Huimin Li, Lingling Liu, Fang Jiao, Licheng Li, Qing |
| contents | Trajectory-controlled video generation has become essential for controllable video generation. While current methods perform well under small-view camera motions, they degrade significantly with large-view motions. Existing solutions for extreme-view synthesis typically require dedicated video pairs, demanding substantial annotation effort. To address these limitations, we propose Dynamic Extreme VIew Synthesis-GRPO (DEVIS-GRPO), a GRPO-based framework for trajectory-controlled video generation, the first online policy gradient method for extreme view video generation. Central to our approach is a novel sampling strategy: Accumulative Dynamic Extreme VIew Synthesis (ADEVIS), which achieves large-view camera motions by progressively accumulating small-view increments. This method delivers two key advantages: 1) enhanced training efficiency, as it eliminates the need to warm-start the policy model by collecting expensive paired large-view videos, and 2) increased sampling diversity, achieved by flexibly varying trajectory configurations. Finally, we designed a multi-level consistency-quality reward function to select high-quality samples for model optimization. Experiments on the Kubric-4D, iPhone, and DL3DV datasets demonstrate our method's superiority. On Kubric-4D, we achieve relative improvements of 21.57% in PSNR and 7.31% in SSIM over the second-best method in non-occlusion areas. On iPhone, LPIPS is reduced by 18.56%. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_16937 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DEVIS-GRPO: Unleashing GRPO on Dynamic Extreme View Synthesis Zuo, Yi Wu, Huimin Li, Lingling Liu, Fang Jiao, Licheng Li, Qing Computer Vision and Pattern Recognition Trajectory-controlled video generation has become essential for controllable video generation. While current methods perform well under small-view camera motions, they degrade significantly with large-view motions. Existing solutions for extreme-view synthesis typically require dedicated video pairs, demanding substantial annotation effort. To address these limitations, we propose Dynamic Extreme VIew Synthesis-GRPO (DEVIS-GRPO), a GRPO-based framework for trajectory-controlled video generation, the first online policy gradient method for extreme view video generation. Central to our approach is a novel sampling strategy: Accumulative Dynamic Extreme VIew Synthesis (ADEVIS), which achieves large-view camera motions by progressively accumulating small-view increments. This method delivers two key advantages: 1) enhanced training efficiency, as it eliminates the need to warm-start the policy model by collecting expensive paired large-view videos, and 2) increased sampling diversity, achieved by flexibly varying trajectory configurations. Finally, we designed a multi-level consistency-quality reward function to select high-quality samples for model optimization. Experiments on the Kubric-4D, iPhone, and DL3DV datasets demonstrate our method's superiority. On Kubric-4D, we achieve relative improvements of 21.57% in PSNR and 7.31% in SSIM over the second-best method in non-occlusion areas. On iPhone, LPIPS is reduced by 18.56%. |
| title | DEVIS-GRPO: Unleashing GRPO on Dynamic Extreme View Synthesis |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.16937 |