Research map / Video / Video generation
MagicMotion: Controllable Video Generation with Dense-to-Sparse Trajectory Guidance
Quanhao Li, Zhen Xing, Rui Wang, Hui Zhang, Qi Dai, Zuxuan Wu
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 |
|---|---|---|---|
| train on MagicData instead of MeViS+MOSE (ablation) | MagicMotion Stage2 · MagicBench | Mask IoU (%): 73.39 -> 76.61 | Improved |
| train on MagicData instead of MeViS+MOSE (ablation) | MagicMotion Stage2 · MagicBench | FID: 24.24 -> 15.17 | Improved |
| use progressive dense-to-sparse training instead of training boxes from scratch (ablation) | MagicMotion Stage2 · DAVIS | Box IoU (%): 67.34 -> 72.84 | Improved |
| use progressive dense-to-sparse training instead of training boxes from scratch (ablation) | MagicMotion Stage2 · MagicBench | Mask IoU (%): 74.61 -> 76.61 | Improved |
| add latent segment loss (ablation) | MagicMotion Stage2 · DAVIS | Mask IoU (%): 49.19 -> 53.94 | Improved |
| add latent segment loss (ablation) | MagicMotion Stage2 · MagicBench | Box IoU (%): 78.02 -> 81.45 | Improved |
| move from mask control (Stage1) to box control (Stage2) | MagicMotion CogVideoX · MagicBench | Mask IoU (%): 91.57 -> 76.61 | Hurt |
| MagicMotion vs best prior method on MagicBench mask accuracy | MagicMotion Stage1 CogVideoX · MagicBench | Mask IoU (%): 68.78 -> 91.57 | Improved |
| MagicMotion vs best prior FVD on MagicBench | MagicMotion Stage2 CogVideoX · MagicBench | FVD: 168.82 -> 107.21 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- Tora — Tora: Trajectory-oriented Diffusion Transformer for Video Generation (CVPR 2025) · better on 3 of 3 shared results
- SG-I2V — SG-I2V: Self-Guided Trajectory Control in Image-to-Video Generation (ICLR 2025) · better on 3 of 3 shared results
- DragAnything — Drag Anything: Motion Control for Anything using Entity Representation (ECCV 2024) · better on 1 of 1 shared result
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 0.90)
- MotionPro: A Precise Motion Controller for Image-to-Video Generation (CVPR 2025 · similarity 0.89)
- Motion Prompting: Controlling Video Generation with Motion Trajectories (CVPR 2025 · similarity 0.88)
- Image Conductor: Precision Control for Interactive Video Synthesis (AAAI 2025 · similarity 0.88)
- 3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation (ICLR 2025 · similarity 0.87)
- TokenMotion: Decoupled Motion Control via Token Disentanglement for Human-centric Video Generation (CVPR 2025 · similarity 0.87)
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.
