Research map / 3D vision / Human pose estimation
Progressively Generating Better Initial Guesses Towards Next Stages for High-Quality Human Motion Prediction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use the four-stage progressive prediction framework rather than one stage (ablation) | PGBIG · Human3.6M | MPJPE: 67.48 -> 65.02 | Helped |
| Add intermediate-stage supervision (ablation) | PGBIG · Human3.6M | MPJPE: 67.07 -> 65.02 | Helped |
| Use recursively smoothed targets rather than ground truth at every stage (ablation) | PGBIG · Human3.6M | MPJPE: 66.11 -> 65.02 | Helped |
| Use the Encoder-Copy-Decoder stage network instead of LTD (ablation) | PGBIG · Human3.6M | MPJPE: 67.15 -> 65.02 | Helped |
| Use S-DGCN and T-DGCN instead of ST-GCN (ablation) | PGBIG · Human3.6M | MPJPE: 67.97 -> 65.02 | Helped |
| Add the Copy operator to the decoder (ablation) | PGBIG · Human3.6M | MPJPE: 65.99 -> 65.02 | Helped |
| Use four stages rather than the tested alternatives (ablation) | PGBIG · Human3.6M | MPJPE | Helped |
| Copy once rather than not copying (ablation) | PGBIG Encoder-Copy-Decoder · Human3.6M | MPJPE: 65.99 -> 65.02 | Helped |
| Copy three times rather than once (ablation) | PGBIG Encoder-Copy-Decoder · Human3.6M | MPJPE: 65.02 -> 65.35 | Hurt |
| Copy along the temporal rather than channel dimension (ablation) | PGBIG Encoder-Copy-Decoder · Human3.6M | MPJPE: 65.75 -> 65.02 | Helped |
| Copy along the temporal rather than spatial dimension (ablation) | PGBIG Encoder-Copy-Decoder · Human3.6M | MPJPE: 65.21 -> 65.02 | Helped |
| Use AAS rather than Gaussian smoothing with window 15 (ablation) | PGBIG intermediate targets · Human3.6M | MPJPE: 66.6 -> 65.0 | Helped |
Closest papers by meaning
- GCNext: Towards the Unity of Graph Convolutions for Human Motion Prediction (AAAI 2024 · similarity 0.88)
- SimpliHuMoN: Simplifying Human Motion Prediction (NeurIPS 2026 · similarity 0.87)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.87)
- MotionMap: Representing Multimodality in Human Pose Forecasting (CVPR 2025 · similarity 0.86)
- ALIEN: Implicit Neural Representations for Human Motion Prediction under Arbitrary Latency (CVPR 2025 · similarity 0.86)
- HVIS: A Human-like Vision and Inference System for Human Motion Prediction (AAAI 2025 · 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.