Research map / Video / Video generation
OmniDrag: Enabling Motion Control for Omnidirectional Image-to-Video Generation
Weiqi Li, Shijie Zhao, Chong Mou, Xuhan Sheng, Zhenyu Zhang, Qian Wang, Junlin Li, Li Zhang, Jian Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add joint fine-tuning of temporal attention layers (ablation) | OmniDrag · ERP image | FID / FVD: 182.72 -> 171.41 FID, 982.41 -> 933.73 FVD | Improved |
| add Move360 dataset to training (vs WEB360 only) (ablation) | OmniDrag · ERP image | ObjMC: 0.327 -> 0.044 | Improved |
| add Move360 dataset to training (vs WEB360 only) (ablation) | OmniDrag · ERP image | FID: 167.56 -> 171.41 | Hurt |
| add HEALPix initialization (ablation) | OmniDrag · ERP image | ObjMC: 0.226 -> 0.044 | Improved |
| add spherical distance filter (ablation) | OmniDrag · ERP image | ObjMC: 0.113 -> 0.044 | Improved |
| add spherical interpolation (ablation) | OmniDrag · ERP image | ObjMC: 0.053 -> 0.044 | Improved |
| add spherical interpolation (ablation) | OmniDrag · Horizontal 8 viewports | FVD: 342.18 -> 322.22 | Improved |
| use OmniDrag instead of DragNUWA | drag-style I2V · ERP image | ObjMC: 0.418 -> 0.044 | Improved |
| use OmniDrag instead of DragNUWA | drag-style I2V · ERP image | FID: 164.84 -> 171.41 | Hurt |
| use OmniDrag instead of DragAnything | drag-style I2V · Human evaluation | overall preference: 9.3% -> 75.7% | Improved |
Later papers that compared with this method
1 later accepted paper reports OmniDrag as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- 360Explorer: Exploring 4D Controllable World in Panoramic Videos (AAAI 2026 · better on 8 of 8 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DragAnything — Drag Anything: Motion Control for Anything using Entity Representation (ECCV 2024) · better on 7 of 7 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance (NeurIPS 2025 · similarity 0.88)
- MotionPro: A Precise Motion Controller for Image-to-Video Generation (CVPR 2025 · similarity 0.87)
- Autoregressive Omni-Aware Outpainting for Open-Vocabulary 360-Degree Image Generation (AAAI 2024 · similarity 0.87)
- OmniLocalRF: Omnidirectional Local Radiance Fields from Dynamic Videos (CVPR 2024 · similarity 0.86)
- SG-I2V: Self-Guided Trajectory Control in Image-to-Video Generation (ICLR 2025 · similarity 0.86)
- IM-Zero: Instance-level Motion Controllable Video Generation in a Zero-shot Manner (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
