Research map / 3D vision / 3D generation
EgoMAN: Interaction-Structured Reasoning for Egocentric 3D Hand Trajectory Prediction
Mingfei Chen, Yifan Wang, Zhengqin Li, Homanga Bharadhwaj, Yujin Chen, Chuan Qin, Ziyi Kou, Yuan Tian, Eric Whitmire, Rajinder Sodhi, Hrvoje Benko, Eli Shlizerman and 1 more
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 6DoF Trajectory-Token Interface waypoints (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | ADE: 0.161 -> 0.151 | Improved |
| add 6DoF Trajectory-Token Interface waypoints (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | DTW: 0.147 -> 0.137 | Improved |
| add 6DoF Trajectory-Token Interface waypoints (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | Rotation error: 35.90 -> 33.88 | Improved |
| add interaction-structured reasoning without waypoints (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | ADE: 0.162 -> 0.161 | No clear effect |
| add Motion Expert pretraining (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | ADE: 0.215 -> 0.150 | Improved |
| add reasoning and Motion Expert pretraining (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | ADE: 0.273 -> 0.151 | Improved |
| use 6DoF waypoints instead of decoder hidden state (Emb) (ablation) | EgoMAN · EgoMAN-Unseen (K=1) ablation | FDE: 0.210 -> 0.206 | No clear effect |
| add interaction-structured reasoning (ablation) | EgoMAN · EgoMAN-Unseen (K=10) | ADE: 0.141 -> 0.124 | Improved |
| add interaction-structured reasoning (full data) (ablation) | EgoMAN · EgoMAN-Unseen data efficiency | ADE: 0.140 -> 0.125 | Improved |
| add interaction-structured reasoning (20% data) (ablation) | EgoMAN · EgoMAN-Unseen data efficiency | ADE: 0.16 -> 0.13 | Improved |
| use Flow Matching Motion Expert instead of MMTwin* | FM-base · EgoMAN-Unseen (K=10) | ADE: 0.206 -> 0.160 | Improved |
| use EgoMAN instead of HandsOnVLM* | EgoMAN · EgoMAN-Unseen (K=10) | ADE: 0.171 -> 0.124 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VidBot — VidBot: Learning Generalizable 3D Actions from In-the-Wild 2D Human Videos for Zero-Shot Robotic Manipulation (CVPR 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Uni-Hand: Universal Hand Motion Forecasting in Egocentric Views (TPAMI 2026 · similarity 0.92)
- Joint Hand Motion and Interaction Hotspots Prediction From Egocentric Videos (CVPR 2022 · similarity 0.92)
- How Do I Do That? Synthesizing 3D Hand Motion and Contacts for Everyday Interactions (CVPR 2025 · similarity 0.90)
- Hand-Centric Motion Refinement for 3D Hand-Object Interaction via Hierarchical Spatial-Temporal Modeling (AAAI 2024 · similarity 0.90)
- MEgoHand: Multimodal Egocentric Hand-Object Interaction Motion Generation (NeurIPS 2025 · similarity 0.89)
- Modeling Fine-Grained Hand-Object Dynamics for Egocentric Video Representation Learning (ICLR 2025 · similarity 0.89)
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.
