Research map / Video / Video generation
MatchDiffusion: Training-free Generation of Match-Cuts
Alejandro Pardo, Fabio Pizzati, Tong Zhang, Alexander Pondaven, Philip Torr, Juan Camilo Perez, Bernard Ghanem
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 MatchDiffusion joint-then-disjoint diffusion instead of independent T2V generation | CogVideoX-5B · Match-cut generation | Motion consistency: 0.40 -> 0.70 | Improved |
| use MatchDiffusion instead of independent T2V generation | CogVideoX-5B · Match-cut generation | LPIPS: 0.74 -> 0.32 | Improved |
| use MatchDiffusion instead of V2V baseline | CogVideoX-5B · Match-cut generation | CLIPScore: 0.31 -> 0.34 | Improved |
| use MatchDiffusion instead of V2V baseline | CogVideoX-5B · Match-cut generation | LPIPS: 0.31 vs 0.32 | No clear effect |
| use MatchDiffusion instead of motion transfer (SMM) | CogVideoX-5B · Match-cut generation | LPIPS: 0.74 -> 0.32 | Improved |
| use MatchDiffusion instead of MOFT | CogVideoX-5B · Match-cut user study | strongly agree share: 12.36% -> 39.44% | Improved |
| increase number of joint diffusion steps K (ablation) | MatchDiffusion · Match-cut generation | CLIPScore | Hurt |
| increase number of joint diffusion steps K (ablation) | MatchDiffusion · Match-cut generation | LPIPS: LPIPS to zero at K=50 | Improved |
| increase classifier-free guidance (ablation) | MatchDiffusion · Match-cut generation | CLIPScore | Hurt |
| apply user intervention tau to joint-diffusion output (ablation) | MatchDiffusion · Match-cut interventions | CLIPScore | No clear effect |
| combine paths with linear decay instead of averaging (ablation) | MatchDiffusion · Match-cut generation | Motion fidelity | Mixed |
Closest papers by meaning
- AdaDiff: Adaptive Step Selection for Fast Diffusion Models (AAAI 2025 · similarity 0.88)
- Emergent Temporal Correspondences from Video Diffusion Transformers (NeurIPS 2025 · similarity 0.88)
- VideoGuide: Improving Video Diffusion Models without Training Through a Teacher's Guide (CVPR 2025 · similarity 0.87)
- SEINE: Short-to-Long Video Diffusion Model for Generative Transition and Prediction (ICLR 2024 · similarity 0.87)
- Video Diffusion Models Are Strong Video Inpainter (AAAI 2025 · similarity 0.87)
- LongDiff: Training-Free Long Video Generation in One Go (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.
