Research map / Video / Video generation
MoCoDiff: A Controllable Autoregressive Diffusion Model for Expressive Motion Generation
Wenfeng Song, Xuehan Wang, Shuai Li, Yi Chen, Yuting Guo, Zhenyu Wu, Xingliang Jin, Chenglizhao Chen, Fei Hou, Hongyu Wu, Aimin Hao
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 autoregressive diffusion with history-aware TIMC (ablation) | MoCoDiff · Long-sequence stylized motion | AUJ: 2.96 -> 1.58 | Improved |
| add autoregressive diffusion with history-aware TIMC (ablation) | MoCoDiff · Long-sequence stylized motion | SRA: 25.69 -> 26.37 | Improved |
| use IMC injection instead of ControlNet duplicated U-Net (ablation) | MoCoDiff · Long-sequence stylized motion | SRA: 15.38 -> 26.37 | Improved |
| use IMC injection instead of ControlNet duplicated U-Net (ablation) | MoCoDiff · Long-sequence stylized motion | AUJ: 2.48 -> 1.58 | Improved |
| freeze the diffusion backbone (ablation) | MoCoDiff · Long-sequence stylized motion | FID: 18.42 -> 5.95 | Improved |
| restrict IMC modulation to the encoder only (ablation) | MoCoDiff · Long-sequence stylized motion | FID: 16.56 -> 5.95 | Improved |
| increase CFG style weight from 1.0 to 1.5 (ablation) | MoCoDiff Style IMC · Single motion generation | SRA / FID: SRA 27.21 -> 32.15; FID 5.56 -> 12.84 | Mixed |
| decrease CFG style weight from 1.0 to 0.5 (ablation) | MoCoDiff Style IMC · Single motion generation | SRA / FID: SRA 27.21 -> 7.82; FID 5.56 -> 1.99 | Mixed |
| lengthen history window beyond 10 frames (ablation) | MoCoDiff · Long-sequence stylized motion | SRA / AUJ | Hurt |
| use MoCoDiff instead of autoregressive-augmented baselines | AutoMLD+SMooDi · Time efficiency | FPS: 28.57 -> 136.89 (vs best baseline AutoMLD+MCM LDM) | Improved |
Closest papers by meaning
- Enhanced Fine-Grained Motion Diffusion for Text-Driven Human Motion Synthesis (AAAI 2024 · similarity 0.87)
- MotionMix: Weakly-Supervised Diffusion for Controllable Motion Generation (AAAI 2024 · similarity 0.86)
- Towards Detailed Text-to-Motion Synthesis via Basic-to-Advanced Hierarchical Diffusion Model (AAAI 2024 · similarity 0.86)
- FlowMo: Variance-Based Flow Guidance for Coherent Motion in Video Generation (NeurIPS 2025 · similarity 0.86)
- AMD: Autoregressive Motion Diffusion (AAAI 2024 · similarity 0.86)
- Arbitrary Motion Style Transfer with Multi-condition Motion Latent Diffusion Model (CVPR 2024 · similarity 0.86)
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.
