Research map / Video / Video-language understanding
CoT-Edit: Let CoT Guide Instruction Video Editing
Sen Liang, Fengbin Guan, Youliang Zhang, Xin Li, Zhibo 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 Qwen3-VL features to the Editor (ablation) | Editor · Koala-36M 100-video eval | physical plausibility: 0.501 -> 0.674 | Improved |
| add Qwen3-VL features to the Editor (ablation) | Editor · Koala-36M 100-video eval | spatial relations: 0.754 -> 0.758 | No clear effect |
| add Qwen3-VL features to the Editor (ablation) | Editor · Koala-36M 100-video eval | instruction following: 0.575 -> 0.598 | Improved |
| add Qwen3-VL features to the Editor (ablation) | Editor · Koala-36M 100-video eval | overall editing quality: 0.598 -> 0.631 | Improved |
| add mask branch with Mask-Connector (ablation) | Editor with MLLM · Koala-36M 100-video eval | spatial relations: 0.758 -> 0.807 | Improved |
| add mask branch with Mask-Connector (ablation) | Editor with MLLM · Koala-36M 100-video eval | physical plausibility: 0.674 -> 0.643 | Hurt |
| add mask branch with Mask-Connector (ablation) | Editor with MLLM · Koala-36M 100-video eval | overall editing quality: 0.631 -> 0.638 | Improved |
| add Reverse-Connector to mask branch (ablation) | Editor + mask branch · Koala-36M 100-video eval | physical plausibility: 0.643 -> 0.681 | Improved |
| add Reverse-Connector to mask branch (ablation) | Editor + mask branch · Koala-36M 100-video eval | spatial relations: 0.807 -> 0.815 | Improved |
| add Reverse-Connector to mask branch (ablation) | Editor + mask branch · Koala-36M 100-video eval | instruction following: 0.586 -> 0.609 | Improved |
| add Reverse-Connector to mask branch (ablation) | Editor + mask branch · Koala-36M 100-video eval | overall editing quality: 0.638 -> 0.647 | Improved |
| add CoT reasoning to the MLLM planner (Qwen3-VL-32B and Gemini 2.5 Pro) (ablation) | planner · Koala-36M 100-video eval | qualitative physical plausibility | Improved |
Closest papers by meaning
- Meta-CoT: Enhancing Granularity and Generalization in Image Editing (CVPR 2026 · similarity 0.90)
- Text-Vision Co-Instructed Image Editing (NeurIPS 2026 · similarity 0.90)
- Learning Action and Reasoning-Centric Image Editing from Videos and Simulation (NeurIPS 2024 · similarity 0.88)
- Guiding Instruction-based Image Editing via Multimodal Large Language Models (ICLR 2024 · similarity 0.87)
- Bernini: Latent Semantic Planning for Video Diffusion (NeurIPS 2026 · similarity 0.87)
- From Scale to Speed: Adaptive Test-Time Scaling for Image Editing (CVPR 2026 · 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.
