Research map / Point clouds / Point cloud classification
FlowCloud: Learning Continuous Spatiotemporal Dynamics from Unpaired Sparse Point Cloud Snapshots
Yinbo Liu, Keyang Ye, Wenshan Sun, Handi Gao, Tian Tian
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 Neural ODE continuous dynamics (vs discrete dynamics) (ablation) | FlowCloud · Simulation 2D extrapolation | Chamfer distance: 0.5517 -> 0.0449 | Improved |
| add Point Swin Transformer encoder (vs PointNet++) (ablation) | FlowCloud · Simulation 2D | cell type accuracy: 0.812 -> 0.898 | Improved |
| add global sequence context (vs recursive aggregation) (ablation) | FlowCloud · Simulation 2D extrapolation | Chamfer distance: 0.3105 -> 0.0449 | Improved |
| add global sequence context (vs recursive aggregation) (ablation) | FlowCloud · Simulation 2D | cell type accuracy: 0.763 -> 0.898 | Improved |
| add trajectory consistency loss (ablation) | FlowCloud · Simulation 2D extrapolation | Chamfer distance: 0.0691 -> 0.0449 | Improved |
| add Transformer attention aggregator (vs mean pooling) (ablation) | FlowCloud · Simulation 2D extrapolation | Chamfer distance: 0.0840 -> 0.0449 | Improved |
| FlowCloud vs best baseline SPCMNet on human motion | FlowCloud · Human 3D interpolation | Chamfer distance: 0.0470 -> 0.0268 | Improved |
| FlowCloud vs OT baseline Moscot on Drosophila cell types | FlowCloud · Drosophila 3D | cell type accuracy: 0.0442 -> 0.1086 | Improved |
| FlowCloud vs SPCMNet on Ambystoma geometry | FlowCloud · Ambystoma 2D | Sinkhorn distance: Meteornet 0.2881 vs FlowCloud 0.2868 | Mixed |
| FlowCloud vs Meteornet on Drosophila Sinkhorn | FlowCloud · Drosophila 3D | Sinkhorn distance: 0.2021 -> 0.2238 | Hurt |
| FlowCloud vs SPCMNet on simulation extrapolation cell type accuracy | FlowCloud · Simulation 2D extrapolation | cell type accuracy: 0.4056 -> 0.4581 | 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.
- Meteornet — MeteorNet: Deep Learning on Dynamic 3D Point Cloud Sequences (ICCV 2019) · 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
- Neural Persistence Dynamics (NeurIPS 2024 · similarity 0.85)
- Modeling Microenvironment Trajectories on Spatial Transcriptomics with NicheFlow (NeurIPS 2025 · similarity 0.85)
- TRENDy: Temporal Regression of Effective Nonlinear Dynamics (ICLR 2025 · similarity 0.85)
- ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow (NeurIPS 2026 · similarity 0.85)
- How to build a consistency model: Learning flow maps via self-distillation (NeurIPS 2025 · similarity 0.85)
- Rectifying Categorical Flows on Statistical Manifolds for One-Step Generation (NeurIPS 2026 · 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.
