Research map / Video / Video segmentation and tracking
Remembering What is Important: A Factorised Multi-Head Retrieval and Auxiliary Memory Stabilisation Scheme for Human Motion Prediction
Tharindu Fernando, Harshala Gammulle, Sridha Sridharan, Simon Denman, Clinton Fookes
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 |
|---|---|---|---|
| use factorised multi-head auxiliary memory model instead of GA-MIN | GA-MIN · H3.6M short-term average 400ms | MPJPE: 52.2 -> 34.8 | Improved |
| use factorised multi-head auxiliary memory model instead of GA-MIN | GA-MIN · H3.6M long-term average 1000ms | MPJPE: 72.16 -> 54.62 | Improved |
| use FMS-AM instead of GA-MIN on long-term walking | GA-MIN · H3.6M walking 1000ms | MPJPE: 42.8 -> 31.6 | Improved |
| use FMS-AM instead of GA-MIN on subject-specific discussion | GA-MIN · H3.6M discussion 1000ms | MPJPE: 106.3 -> 79.2 | Improved |
| use FMS-AM instead of GA-MIN on directions | GA-MIN · H3.6M directions 1000ms | MPJPE: 100.0 -> 74.8 | Improved |
| use FMS-AM instead of GA-MIN on jumping | GA-MIN · CMU-Mocap jumping 400ms | MPJPE: 91.1 -> 79.2 | Improved |
| use FMS-AM instead of GA-MIN on soccer | GA-MIN · CMU-Mocap soccer 400ms | MPJPE: 49.4 -> 31.4 | Improved |
| use FMS-AM instead of GA-MIN on running | GA-MIN · CMU-Mocap running 400ms | MPJPE: 26.1 -> 18.8 | Improved |
| use FMS-AM instead of GA-MIN on directing traffic at 80ms | GA-MIN · CMU-Mocap directing traffic 80ms | MPJPE: 5.7 -> 5.8 | No clear effect |
| use FMS-AM instead of DPnet on washwindow at 160ms | DPnet · CMU-Mocap washwindow 160ms | MPJPE: 9.8 -> 10.0 | No clear effect |
| use FMS-AM instead of GA-MIN on short-term average at 320ms | GA-MIN · CMU-Mocap | MPJPE | Improved |
Closest papers by meaning
- Multi-level Motion Attention for Human Motion Prediction (IJCV 2021 · similarity 0.88)
- History Repeats Itself: Human Motion Prediction via Motion Attention (ECCV 2020 · similarity 0.88)
- Investigating Pose Representations and Motion Contexts Modeling for 3D Motion Prediction (TPAMI 2022 · similarity 0.87)
- Overlooked Poses Actually Make Sense: Distilling Privileged Knowledge for Human Motion Prediction (ECCV 2022 · similarity 0.87)
- ALIEN: Implicit Neural Representations for Human Motion Prediction under Arbitrary Latency (CVPR 2025 · similarity 0.86)
- HMVLM:Human Motion-Vision-Language Model via MoE LoRA (NeurIPS 2025 · similarity 0.85)
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.
