Research map / Video / Video generation
Learning Uncoupled-Modulation CVAE for 3D Action-Conditioned Human Motion Synthesis
Chongyang Zhong, Lei Hu, Zihao Zhang, Shihong Xia
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 action-aware modulation (FiLM) (ablation) | UM-CVAE · UESTC | FIDte: 25.17 -> 8.58 | Improved |
| add action-aware modulation (FiLM) (ablation) | UM-CVAE · UESTC | accuracy: 88.9 -> 93.0 | Improved |
| add action-aware modulation (FiLM) (ablation) | UM-CVAE · HumanAct12 | accuracy: 92.7 -> 95.8 | Improved |
| add instance normalization to encoder (ablation) | UM-CVAE · UESTC | FIDte: 16.42 -> 8.58 | Improved |
| add instance normalization to encoder (ablation) | UM-CVAE · UESTC | accuracy: 91.6 -> 93.0 | Improved |
| add instance normalization to encoder (ablation) | UM-CVAE · HumanAct12 | accuracy: 93.9 -> 95.8 | Improved |
| add latent reinforcement (ablation) | UM-CVAE · UESTC | diversity: 26.78 -> 31.85 | Improved |
| add latent reinforcement (ablation) | UM-CVAE · UESTC | FIDte: 12.71 -> 8.58 | Improved |
| add latent reinforcement (ablation) | UM-CVAE · HumanAct12 | diversity: 5.18 -> 6.81 | Improved |
| augment BABEL-60 training set with generated motions | UM-CVAE · BABEL-60 | Top-1 recognition accuracy: 33.41 -> 35.14 | Improved |
| use UM-CVAE instead of ACTOR | UM-CVAE · UESTC | FIDte: 18.95 -> 8.58 | Improved |
| use UM-CVAE instead of ACTOR | UM-CVAE · BABEL-60 | Top-1 recognition accuracy: 25.37 -> 29.04 | Improved |
Closest papers by meaning
- A Unified 3D Human Motion Synthesis Model via Conditional Variational Auto-Encoder (ICCV 2021 · similarity 0.93)
- Action-Conditioned 3D Human Motion Synthesis With Transformer VAE (ICCV 2021 · similarity 0.91)
- Seamless Human Motion Composition with Blended Positional Encodings (CVPR 2024 · similarity 0.88)
- Contextually Plausible and Diverse 3D Human Motion Prediction (ICCV 2021 · similarity 0.88)
- A Unified Masked Autoencoder with Patchified Skeletons for Motion Synthesis (AAAI 2024 · similarity 0.87)
- Everything2Motion: Synchronizing Diverse Inputs via a Unified Framework for Human Motion Synthesis (AAAI 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.
