Research map / Video / Video generation
DIVE: Taming DINO for Subject-Driven Video Editing
Yi Huang, Wei Xiong, He Zhang, Chaoqi Chen, Jianzhuang Liu, Mingfu Yan, Shifeng Chen
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 DINO motion guidance instead of fine-tuned motion module only (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Image alignment: 67.49 -> 84.27 | Improved |
| add DINO motion guidance instead of fine-tuned motion module only (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Overall video quality: 0.737 -> 0.775 | Improved |
| add DINO motion guidance instead of fine-tuned motion module only (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Text alignment: 29.91 -> 29.43 | Hurt |
| raise guidance weight lambda from 0 to 1.0 (frozen motion module) (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Temporal consistency: 87.24 -> 92.33 | Improved |
| raise guidance weight lambda from 0.3 to 1.0 (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Overall video quality: 0.628 -> 0.775 | Improved |
| raise guidance weight lambda from 0 to 1.0 (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Overall video quality: 0.581 -> 0.775 | Improved |
| add DINO features to identity registration (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Image alignment: 77.36 -> 84.27 | Improved |
| add DINO features to identity registration (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Overall video quality: 0.722 -> 0.775 | Improved |
| add DINO features to identity registration (ablation) | DIVE · Reference-image-guided subject editing (30 videos) | Temporal consistency: 92.12 -> 92.33 | No clear effect |
| use DIVE instead of FLATTEN (best baseline) for reference-image editing | DIVE · Reference-image-guided subject editing (30 videos) | Overall video quality: 0.683 -> 0.775 | Improved |
| use DIVE instead of AnyV2V for reference-image editing | DIVE · Reference-image-guided subject editing (30 videos) | Image alignment: 78.26 -> 84.27 | Improved |
| use DIVE instead of RAVE for text-guided editing | DIVE · Text-guided subject editing (30 videos) | Temporal consistency: 95.12 -> 95.89 | Improved |
Later papers that compared with this method
1 later accepted paper reports DiVE as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- RLGF: Reinforcement Learning with Geometric Feedback for Autonomous Driving Video Generation (NeurIPS 2025 · better on 6 of 6 shared results)
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.
- RAVE — RAVE: Randomized Noise Shuffling for Fast and Consistent Video Editing with Diffusion Models (CVPR 2024) · better on 7 of 7 shared results
- FLATTEN — FLATTEN: optical FLow-guided ATTENtion for consistent text-to-video editing (ICLR 2024) · better on 7 of 7 shared results
- Slicedit — Slicedit: Zero-Shot Video Editing With Text-to-Image Diffusion Models Using Spatio-Temporal Slices (ICML 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
- MotionEditor: Editing Video Motion via Content-Aware Diffusion (CVPR 2024 · similarity 0.87)
- COVE: Unleashing the Diffusion Feature Correspondence for Consistent Video Editing (NeurIPS 2024 · similarity 0.86)
- Señorita-2M: A High-Quality Instruction-based Dataset for General Video Editing by Video Specialists (NeurIPS 2025 · similarity 0.85)
- VideoDirector: Precise Video Editing via Text-to-Video Models (CVPR 2025 · similarity 0.85)
- ConMo: Controllable Motion Disentanglement and Recomposition for Zero-Shot Motion Transfer (CVPR 2025 · similarity 0.85)
- VideoSwap: Customized Video Subject Swapping with Interactive Semantic Point Correspondence (CVPR 2024 · similarity 0.85)
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.
