Research map / 3D vision / Human pose estimation
PoseScript: Linking 3D Human Poses and Natural Language
Ginger Delmas, Philippe Weinzaepfel, Thomas Lucas, Francesc Moreno-Noguer, Grégory Rogez
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 |
|---|---|---|---|
| pretrain retrieval model on automatic captions before finetuning on human captions (ablation) | text-to-pose retrieval (GloVe-biGRU) · PoseScript-H test | mRecall: 23.0 -> 40.9 | Improved |
| use transformer text encoder instead of GloVe-biGRU (ablation) | retrieval pretrain+FT · PoseScript-H test | mRecall: 40.9 -> 43.3 | Improved |
| add mirroring augmentation (ablation) | retrieval with transformer encoder · PoseScript-H test | mRecall: 43.3 -> 45.3 | Improved |
| train only on automatic captions and test on human captions | retrieval · PoseScript-H test | mRecall: 72.8 -> 5.9 | Hurt |
| add L_reg KL regularization to generative model (ablation) | text-conditioned pose VAE · PoseScript-A test | FID: 0.12 -> 0.07 | Improved |
| add L_reg to generative model (ablation) | text-conditioned pose VAE · PoseScript-A test | mRecall G/R: 41.5 -> 40.0 | No clear effect |
| pretrain generative model on PoseScript-A (ablation) | text-conditioned pose VAE · PoseScript-H test | FID: 0.29 -> 0.04 | Improved |
| pretrain generative model on PoseScript-A (ablation) | text-conditioned pose VAE · PoseScript-H test | mRecall R/G: 5.2 -> 19.5 | Improved |
| use transformer text encoder in generative model (ablation) | text-conditioned pose VAE · PoseScript-H test | mRecall R/G: 19.5 -> 35.6 | Improved |
| add mirroring augmentation to generative model (ablation) | text-conditioned pose VAE · PoseScript-H test | mRecall R/G: 35.6 -> 37.5 | Improved |
| pretrain captioning model on automatic captions (ablation) | pose description generator · PoseScript-H test | R-Precision R@1: 24.39 -> 88.09 | Improved |
| add mirroring augmentation to captioning model (ablation) | pose description generator · PoseScript-H test | BLEU-4: 13.07 -> 13.22 | Improved |
Closest papers by meaning
- ChatPose: Chatting about 3D Human Pose (CVPR 2024 · similarity 0.90)
- UniPose: A Unified Multimodal Framework for Human Pose Comprehension, Generation and Editing (CVPR 2025 · similarity 0.88)
- PoseGen: Learning to Generate 3D Human Pose Dataset with NeRF (AAAI 2024 · similarity 0.88)
- PoseLLaVA: Pose Centric Multimodal LLM for Fine-Grained 3D Pose Manipulation (AAAI 2025 · similarity 0.87)
- FlexPose: Pose Distribution Adaptation with Limited Guidance (AAAI 2025 · similarity 0.87)
- InteractAnything: Zero-shot Human Object Interaction Synthesis via LLM Feedback and Object Affordance Parsing (CVPR 2025 · 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.
