Research map / Video / Video generation
Make It Move: Controllable Image-to-Video Generation With Text Descriptions
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 cross-attention instead of concatenation to align image and text in the motion anchor (ablation) | MAGE · CATER-GEN-v1 | FID: 63.4 -> 62.0 | Helped |
| Use cross-attention instead of concatenation to align image and text in the motion anchor (ablation) | MAGE · CATER-GEN-v1 | LPIPS: 0.20 -> 0.20 | No clear effect |
| Use cross-attention instead of concatenation to align image and text in the motion anchor (ablation) | MAGE · CATER-GEN-v1 | FVD: 30.4 -> 30.5 | No clear effect |
| Use axial instead of vanilla transformers (ablation) | MAGE · CATER-GEN-v1 | GFlops: 96.8 -> 52.7 | Helped |
| Use axial instead of vanilla transformers (ablation) | MAGE · CATER-GEN-v1 | Throughput: 11.5 -> 37.2 | Helped |
| Use axial instead of vanilla transformers (ablation) | MAGE · CATER-GEN-v1 | FID: 62.0 -> 62.7 | Hurt |
| Enable implicit randomness for ambiguous descriptions | MAGE · CATER-GEN-v1 | DIV I3D: 0 -> 0.45 | Helped |
| Enable implicit randomness for ambiguous descriptions | MAGE · CATER-GEN-v2 | DIV I3D: 0 -> 2.06 | Helped |
| Enable implicit randomness for ambiguous descriptions | MAGE · CATER-GEN-v1 | FVD: 31.70 -> 45.49 | Hurt |
| Enable implicit randomness for ambiguous descriptions | MAGE · CATER-GEN-v2 | FVD: 57.55 -> 69.44 | Hurt |
Closest papers by meaning
- TIV-Diffusion: Towards Object-Centric Movement for Text-driven Image to Video Generation (AAAI 2025 · similarity 0.90)
- MotiF: Making Text Count in Image Animation with Motion Focal Loss (CVPR 2025 · similarity 0.90)
- Enhancing Motion in Text-to-Video Generation with Decomposed Encoding and Conditioning (NeurIPS 2024 · similarity 0.90)
- IM-Zero: Instance-level Motion Controllable Video Generation in a Zero-shot Manner (CVPR 2025 · similarity 0.88)
- Motion Prompting: Controlling Video Generation with Motion Trajectories (CVPR 2025 · similarity 0.87)
- ControlVideo: Training-free Controllable Text-to-video Generation (ICLR 2024 · 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.