Research map / Video / Video generation
Hierarchical Enhancement of Semantic Priors for Disentangled Text-Driven Motion Generation
Wenhan Lv, Shaopan Wang, Xiangyu Wu, Tianchu Hang, Zhongquan Jian, Qingqiang Wu
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 Dynamic Cross-Modal Memory (DCMM) (ablation) | HESP · HumanML3D test | FID: 0.167 -> 0.049 | Improved |
| add Hierarchical Cross-Modal Attention (HCA) instead of self-attention (ablation) | HESP · HumanML3D test | FID: 0.116 -> 0.049 | Improved |
| add DCMM and HCA together (ablation) | HESP · HumanML3D test | R-Precision Top-3: 0.826 -> 0.871 | Improved |
| add DCMM (ablation) | HESP · HumanML3D test | R-Precision Top-3: 0.847 -> 0.871 | Improved |
| add HCA (ablation) | HESP · HumanML3D test | R-Precision Top-3: 0.839 -> 0.871 | Improved |
| replace standard VAE with AG-VAE (ablation) | HESP · HumanML3D test latent space | Calinski-Harabasz: 11.95 -> 48.81 | Improved |
| replace standard VAE with AG-VAE (ablation) | HESP · HumanML3D test latent space | Davies-Bouldin: 7.83 -> 5.39 | Improved |
| replace SALAD VAE with AG-VAE (ablation) | HESP · HumanML3D test VAE reconstruction | MAE: 0.450 -> 0.344 | Improved |
| HESP vs SALAD on R-Precision Top-1 | HESP · HumanML3D test | R-Precision Top-1: 0.581 -> 0.600 | Improved |
| HESP vs SALAD on FID | HESP · HumanML3D test | FID: 0.076 -> 0.045 | Improved |
| HESP vs SALAD on FID | HESP · KIT-ML test | FID: 0.296 -> 0.267 | Improved |
Closest papers by meaning
- MotionGPT3: Human Motion as a Second Modality (ICLR 2026 · similarity 0.91)
- MoLingo: Motion-Language Alignment for Text-to-Human Motion Generation (CVPR 2026 · similarity 0.90)
- MotionHiFlow: Text-to-Motion via Hierarchical Flow Matching (CVPR 2026 · similarity 0.90)
- Towards Detailed Text-to-Motion Synthesis via Basic-to-Advanced Hierarchical Diffusion Model (AAAI 2024 · similarity 0.89)
- Move as You Say Interact as You Can: Language-guided Human Motion Generation with Scene Affordance (CVPR 2024 · similarity 0.88)
- Rethinking Diffusion for Text-Driven Human Motion Generation: Redundant Representations, Evaluation, and Masked Autoregression (CVPR 2025 · similarity 0.88)
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.
