Research map / Video / Video generation
MotionShot: Adaptive Motion Transfer across Arbitrary Objects for Text-to-Video Generation
Yanchen Liu, Yanan Sun, Zhening Xing, Junyao Gao, Kai Chen, Wenjie Pei
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 |
|---|---|---|---|
| use MotionShot instead of MotionClone | MotionShot · Motion transfer, 40 reference videos | CLIP text alignment: 26.41 -> 26.95 | Improved |
| use MotionShot instead of MotionClone | MotionShot · Motion transfer, 40 reference videos | CLIP temporal consistency: 97.48 -> 97.81 | Improved |
| use MotionShot instead of MotionClone | MotionShot · Motion transfer, 40 reference videos | user study motion preservation: 2.90 -> 4.95 | Improved |
| use MotionShot instead of MotionClone | MotionShot · Motion transfer, 40 reference videos | user study appearance diversity: 2.50 -> 4.95 | Improved |
| use MotionShot instead of Tune-A-Video | MotionShot · Motion transfer, 40 reference videos | user study text alignment: 2.88 -> 4.94 | Improved |
| use MotionShot instead of Control-A-Video | MotionShot · Motion transfer, 40 reference videos | user study temporal consistency: 2.92 -> 4.90 | Improved |
| fuse SD and DINO features for keypoint matching instead of either alone or X-Pose detector (ablation) | MotionShot · Motion transfer, 40 reference videos | qualitative motion alignment | Improved |
| use TPS-based keypoint shape retargeting instead of original or resized reference sequences (ablation) | MotionShot · Motion transfer, 40 reference videos | qualitative motion-shape alignment | Improved |
| use m=30 sampled keypoints instead of m=10 (ablation) | MotionShot · Motion transfer, 40 reference videos | qualitative TPS deformation | Improved |
| use m=30 sampled keypoints instead of m=60 (ablation) | MotionShot · Motion transfer, 40 reference videos | qualitative TPS deformation | Improved |
| add fake target object generation conditioned on a degraded reference first-frame segmentation | MotionShot · Motion transfer, 40 reference videos | motion transfer stability | Improved |
| use point-based target keypoint guidance instead of TPS-warped frames | MotionShot · Motion transfer, 40 reference videos | temporal attention guidance | Hurt |
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.
- MotionClone — MotionClone: Training-Free Motion Cloning for Controllable Video Generation (ICLR 2025) · better on 6 of 6 shared results
- Tune-A-Video — Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation (ICCV 2023) · better on 6 of 6 shared results
- VideoComposer — VideoComposer: Compositional Video Synthesis with Motion Controllability (NeurIPS 2023) · better on 5 of 5 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
- Space-Time Diffusion Features for Zero-Shot Text-Driven Motion Transfer (CVPR 2024 · similarity 0.89)
- Enhancing Motion in Text-to-Video Generation with Decomposed Encoding and Conditioning (NeurIPS 2024 · similarity 0.88)
- TIV-Diffusion: Towards Object-Centric Movement for Text-driven Image to Video Generation (AAAI 2025 · similarity 0.88)
- SimMotionEdit: Text-Based Human Motion Editing with Motion Similarity Prediction (CVPR 2025 · similarity 0.88)
- MotionBooth: Motion-Aware Customized Text-to-Video Generation (NeurIPS 2024 · similarity 0.87)
- ConMo: Controllable Motion Disentanglement and Recomposition for Zero-Shot Motion Transfer (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.
