Research map / 3D vision / 3D generation
Scene-aware human motion synthesis
Generates or predicts human motion conditioned on 3D scenes, text and gaze.
82 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in 3D generation, keyword cluster “motion · interaction · human”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
20 papers in 2023–24 and 39 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 4.4% to 4.4%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Multi-View People Tracking via Hierarchical Trajectory Composition (CVPR 2016 · 175 citations)
- 2017 · Binge Watching: Scaling Affordance Learning From Sitcoms (CVPR 2017 · 81 citations)
- 2018 · Egocentric Basketball Motion Planning From a Single First-Person Image (CVPR 2018 · 19 citations)
- 2019 · Towards Natural and Accurate Future Motion Prediction of Humans and Animals (CVPR 2019 · 149 citations)
- 2020 · The Garden of Forking Paths: Towards Multi-Future Trajectory Prediction (CVPR 2020 · 182 citations)
- 2021 · Introvert: Human Trajectory Prediction via Conditional 3D Attention (CVPR 2021 · 75 citations)
- 2022 · Generating Diverse and Natural 3D Human Motions From Text (CVPR 2022 · 553 citations)
- 2023 · CIRCLE: Capture in Rich Contextual Environments (CVPR 2023 · 44 citations)
- 2024 · Move as You Say Interact as You Can: Language-guided Human Motion Generation with Scene Affordance (CVPR 2024 · 44 citations)
- 2025 · Nonisotropic Gaussian Diffusion for Realistic 3D Human Motion Prediction (CVPR 2025 · 7 citations)
- 2026 · Understanding Dynamic Scenes in Ego Centric 4D Point Clouds (AAAI 2026 · 2 citations)
Earliest papers here
- Multi-View People Tracking via Hierarchical Trajectory Composition (CVPR 2016 · 175 citations)
- Inferring Forces and Learning Human Utilities From Videos (CVPR 2016 · 83 citations)
- Learning Action Maps of Large Environments via First-Person Vision (CVPR 2016 · 35 citations)
What has been tried
51 papers here report 504 measured changes (a component added, removed or swapped, compared with a baseline): helped 79%, no clear effect 4%, hurt 14%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| use Perceiver instead of MLP for ADM | HUMANISE affordance generation · all dist. | 1.513˘.008 -> 1.086˘.007 | Helped | Move as You Say Interact as You Can: Language-guided Human Motion Generation with Scene Affordance (CVPR 2024 · 44 citations) |
| add data augmentation | TRUMANS static setting · Penemean | 2.010 -> 1.820 | Helped | Scaling Up Dynamic Human-Scene Interaction Modeling (CVPR 2024 · 40 citations) |
| Add persona instructions | Long-term social evolution · Consistency | Helped | Digital Life Project: Autonomous 3D Characters with Social Intelligence (CVPR 2024 · 28 citations) | |
| use compact egocentric sensing instead of egocentric depth images | Replica · SR (%) | 8 -> 92 | Helped | EgoGen: An Egocentric Synthetic Data Generator (CVPR 2024 · 14 citations) |
| add target-object localization | HUMANISE test · goal dist. | 0.592 -> 0.384 | Helped | Generating Human Motion in 3D Scenes from Text Descriptions (CVPR 2024 · 11 citations) |
| Add the Tail Prediction (TP) pretraining task | Argoverse 1 validation · b-minFDE6 | Helped | SEPT: Towards Efficient Scene Representation Learning for Motion Prediction (ICLR 2024 · 10 citations) | |
| add scene information as an input | CIRCLE · MPJPE | 26.2 -> 22.7 | Helped | A Unified Diffusion Framework for Scene-aware Human Motion Estimation from Sparse Signals (CVPR 2024 · 8 citations) |
| Add Typed-Graph Attention layers | AMASS · performance metrics | Helped | Nonisotropic Gaussian Diffusion for Realistic 3D Human Motion Prediction (CVPR 2025 · 7 citations) | |
| use a mixture of proximal and distal sensor pairs instead of only proximal pairs | Mixamo+ · Contact Error | 0.800 -> 0.772 | Helped | Skinned Motion Retargeting with Dense Geometric Interaction Perception (NeurIPS 2024 · 7 citations) |
| Add M3Act walk and run groups to MOTRv2 training data | DanceTrack · HOTA | 68.8 -> 70.1 | Helped | Learning from Synthetic Human Group Activities (CVPR 2024 · 6 citations) |
| Add body attention loss | DCM test set · LCD | 0.310 -> 0.147 | Helped | DanceCamera3D: 3D Camera Movement Synthesis with Music and Dance (CVPR 2024 · 5 citations) |
| Add anchor determination | Camera trajectory evaluation · Translation MSE | 16.2 -> 5.3 | Helped | ChatCam: Empowering Camera Control through Conversational AI (NeurIPS 2024 · 3 citations) |
What reviewers will expect
Datasets and baselines named in the 18 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- AMASS · 3 of 18 papers
Baselines
- HumanMAC · 2 of 18 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 59 papers; from OpenAlex and the papers' affiliations).
- Tsinghua University · 3 papers
- Peking University · 3 papers
- Zhejiang University · 3 papers
- Boston University · 2 papers
- Harbin Institute of Technology · 2 papers
- Singapore University of Technology and Design · 2 papers
- Seoul National University · 2 papers
- NVIDIA · 2 papers
- Carnegie Mellon University · 2 papers
Use it
Download BibTeX of 37 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Monocular 3D object reconstruction (94 papers)
- Hand-object interaction generation (116 papers)
- Learning physics from video (74 papers)
Elsewhere in the topic
- Learning 3D reconstruction from images (196 papers)
- Diffusion-based novel view synthesis (156 papers)
- Semantic scene completion (153 papers)
- Generative indoor scene synthesis (131 papers)
- Feed-forward 3D asset generation and editing (116 papers)
- Score distillation for text-to-3D (106 papers)
- Open-vocabulary 3D scene understanding (95 papers)
- 2D-to-3D pretrained representations (94 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).