Research map / 3D vision / Human pose estimation
Human Motion Analysis with Deep Metric Learning
Huseyin Coskun, David Joseph Tan, Sailesh Conjeti, Nassir Navab, Federico Tombari
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 self-attention (ablation) | MMD-NCA · CMU mocap action recognition (unseen categories) | FPR at 70% TPR: 30.04 -> 20.29 | Improved |
| add self-attention (ablation) | MMD-NCA · Human3.6M action recognition (unseen categories) | FPR at 70% TPR: 41.01 -> 33.13 | Improved |
| add self-attention (ablation) | MMD-NCA · CMU mocap person identification | FPR at 70% TPR: 16.78 -> 2.55 | Improved |
| add layer normalization to the LSTM (ablation) | MMD-NCA · CMU mocap action recognition (unseen categories) | FPR at 70% TPR: 27.95 -> 20.29 | Improved |
| add layer normalization to the LSTM (ablation) | MMD-NCA · Human3.6M action recognition (unseen categories) | FPR at 70% TPR: 38.09 -> 33.13 | Improved |
| add layer normalization to the LSTM (ablation) | MMD-NCA · CMU mocap person identification | FPR at 70% TPR: 7.36 -> 2.55 | Improved |
| use Gaussian mixture kernel instead of linear kernel in MMD (ablation) | MMD-NCA · CMU mocap action recognition (unseen categories) | FPR at 70% TPR: 29.00 -> 20.29 | Improved |
| use Gaussian mixture kernel instead of polynomial kernel in MMD (ablation) | MMD-NCA · CMU mocap person identification | FPR at 70% TPR: 5.93 -> 2.55 | Improved |
| use Gaussian mixture kernel instead of linear kernel in MMD (ablation) | MMD-NCA · CMU mocap person identification | FPR at 95% TPR: 35.75 -> 19.31 | Improved |
| use MMD-NCA loss instead of triplet loss | attentive BiLSTM · CMU mocap action recognition (unseen categories) | FPR at 70% TPR: 28.77 -> 20.29 | Improved |
| use MMD-NCA loss instead of N-Pair loss | attentive BiLSTM · Human3.6M action recognition (unseen categories) | FPR at 70% TPR: 36.52 -> 33.13 | Improved |
| use MMD-NCA loss instead of N-Pair loss | attentive BiLSTM · CMU mocap person identification | FPR at 70% TPR: 4.56 -> 2.55 | Improved |
Closest papers by meaning
- MotionBERT: A Unified Perspective on Learning Human Motion Representations (ICCV 2023 · similarity 0.87)
- Investigating Pose Representations and Motion Contexts Modeling for 3D Motion Prediction (TPAMI 2022 · similarity 0.86)
- GaitSnippet: Gait Recognition Beyond Unordered Sets and Ordered Sequences (ICLR 2026 · similarity 0.86)
- H-MoRe: Learning Human-centric Motion Representation for Action Analysis (CVPR 2025 · similarity 0.86)
- Learning Visual Prompt for Gait Recognition (CVPR 2024 · similarity 0.85)
- Exploring Vision Transformers for 3D Human Motion-Language Models with Motion Patches (CVPR 2024 · 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.
