Research map / Video / Video generation
TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Texts
Chuan Guo, Xinxin Zuo, Sen Wang, Li Cheng
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 motion tokens (ablation) | motion2text model · HumanML3D test (motion-to-text) | Cider: 68.3 -> 72.5 | Improved |
| add motion tokens (ablation) | motion2text model · KIT-ML test (motion-to-text) | Cider: 60.1 -> 79.5 | Improved |
| add motion tokens (ablation) | motion2text model · KIT-ML test (motion-to-text) | R-precision top-1: 0.284 -> 0.359 | Improved |
| add motion tokens (ablation) | motion2text model · HumanML3D test (motion-to-text) | R-precision top-1: 0.483 -> 0.516 | Improved |
| add inverse alignment (GRU model) (ablation) | text2motion · HumanML3D test (text-to-motion) | R-precision top-1: 0.351 -> 0.424 | Improved |
| add inverse alignment (GRU model) (ablation) | text2motion · HumanML3D test (text-to-motion) | R-precision top-3: 0.635 -> 0.729 | Improved |
| add inverse alignment (GRU model) (ablation) | text2motion · HumanML3D test (text-to-motion) | FID: 1.739 -> 1.501 | Improved |
| add inverse alignment (GRU model) (ablation) | text2motion · KIT-ML test (text-to-motion) | FID: 4.814 -> 3.599 | Improved |
| add inverse alignment (GRU model) (ablation) | text2motion · KIT-ML test (text-to-motion) | R-precision top-1: 0.251 -> 0.280 | Improved |
| add inverse alignment (GRU model) (ablation) | text2motion · HumanML3D test (text-to-motion) | MModality: 3.139 -> 2.424 | Hurt |
| add inverse alignment (GRU model) (ablation) | text2motion · KIT-ML test (text-to-motion) | MModality: 4.486 -> 3.292 | Hurt |
| use Transformer instead of GRU in text2motion without inverse alignment | text2motion baseline · HumanML3D test (text-to-motion) | FID: 1.739 -> 1.669 | Mixed |
Closest papers by meaning
- Generating Diverse and Natural 3D Human Motions From Text (CVPR 2022 · similarity 0.93)
- Text Motion Translator: A Bi-Directional Model for Enhanced 3D Human Motion Generation from Open-Vocabulary Descriptions (ECCV 2024 · similarity 0.92)
- OpenT2M: No-frill Motion Generation with Open-source, Large-scale, High-quality Data (CVPR 2026 · similarity 0.91)
- Light-T2M: A Lightweight and Fast Model for Text-to-motion Generation (AAAI 2025 · similarity 0.89)
- MMM: Generative Masked Motion Model (CVPR 2024 · similarity 0.89)
- ReMoGPT: Part-Level Retrieval-Augmented Motion-Language Models (AAAI 2025 · similarity 0.89)
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.
