Research map / Video / Video generation
MIMIC: Mask-Injected Manipulation Video Generation with Interaction Control
Tianxiao Chen, Jintao Rong, Huajin Chen, Jingya Wang, Tao Zhou, Jiming Chen, Qi Ye
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 Pair Prompt Control in Stage II (ablation) | MIMIC · Manipulation video generation (reference-based) | Interaction Rationality (MLLM): 3.8789 -> 4.1381 | Improved |
| add Pair Prompt Control in Stage II (ablation) | MIMIC · Manipulation video generation (reference-based) | Semantic Similarity (MLLM): 2.7526 -> 2.9127 | Improved |
| add IMA attention module in Stage I (ablation) | MIMIC · Manipulation video generation (reference-based) | Interaction Rationality (MLLM): 3.6216 -> 4.1381 | Improved |
| add IMA attention module in Stage I (ablation) | MIMIC · Manipulation video generation (reference-based) | Text Alignment: 0.2548 -> 0.2721 | Improved |
| use two-stage (mask then video) instead of one-stage generation (ablation) | MIMIC · Manipulation video generation (reference-based) | Appearance Consistency: 0.8709 -> 0.9084 | Improved |
| use full image rather than masks only as Stage I condition (ablation) | MIMIC · Manipulation video generation (reference-based) | Interaction Rationality (MLLM): 3.5689 -> 4.1381 | Improved |
| use full image rather than masks only as Stage I condition (ablation) | MIMIC · Manipulation video generation (reference-based) | Subject Consistency: 0.8745 -> 0.9291 | Improved |
| add Adaptive Region Loss (lambda 1.00) over original diffusion loss (ablation) | MIMIC Stage II · SSv2 10k subset, hand interaction (Adaptive Region Loss) | Appearance Consistency: 0.8927 -> 0.9054 | Improved |
| add first-frame mask to Adaptive Region Loss (ablation) | MIMIC Stage II · SSv2 10k subset, hand interaction (Adaptive Region Loss) | Appearance Consistency: 0.9042 -> 0.9054 | No clear effect |
| set Adaptive Region Loss weight lambda to 0.50 (ablation) | MIMIC Stage II · SSv2 10k subset, hand interaction (Adaptive Region Loss) | Subject Consistency: 0.8894 -> 0.8709 | Hurt |
| MIMIC vs strongest baseline FlexiAct on MLLM interaction rationality | MIMIC · Manipulation video generation (reference-based) | Interaction Rationality (MLLM): 3.5529 -> 4.1381 | Improved |
| MIMIC vs MotionClone on text alignment | MIMIC · Manipulation video generation (reference-based) | Text Alignment: 0.2947 -> 0.2721 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MotionClone — MotionClone: Training-Free Motion Cloning for Controllable Video Generation (ICLR 2025) · better on 2 of 2 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
- Mask2IV: Interaction-Centric Video Generation via Mask Trajectories (AAAI 2026 · similarity 0.90)
- Robotic Manipulation by Imitating Generated Videos Without Physical Demonstrations (ICLR 2026 · similarity 0.90)
- Learning Video Generation for Robotic Manipulation with Collaborative Trajectory Control (ICLR 2026 · similarity 0.89)
- SViMo: Synchronized Diffusion for Video and Motion Generation in Hand-object Interaction Scenarios (NeurIPS 2025 · similarity 0.88)
- Generate Subgoal Images before Act: Unlocking the Chain-of-Thought Reasoning in Diffusion Model for Robot Manipulation with Multimodal Prompts (CVPR 2024 · similarity 0.87)
- LongScape: Advancing Long-Horizon Embodied World Models with Context-Aware MoE (NeurIPS 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.
