ICLR 2025 accepted papers on Point clouds
25 of the 3,703 papers accepted at ICLR 2025 (0.7%) are on Point clouds, compared with 0.8% at ICLR 2024. They fall into 4 topics.
See ICLR 2025 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Point cloud classification · 11 papers
- E(3)-equivariant models cannot learn chirality: Field-based molecular generation (1 citation)
- GS-LiDAR: Generating Realistic LiDAR Point Clouds with Panoramic Gaussian Splatting (1 citation)
- On the Completeness of Invariant Geometric Deep Learning Models (1 citation)
- Learning Equivariant Non-Local Electron Density Functionals
- PvNeXt: Rethinking Network Design and Temporal Motion for Point Cloud Video Recognition
- UniGS: Unified Language-Image-3D Pretraining with Gaussian Splatting
- X-Drive: Cross-modality Consistent Multi-Sensor Data Synthesis for Driving Scenarios
- Beyond Canonicalization: How Tensorial Messages Improve Equivariant Message Passing
- CityAnchor: City-scale 3D Visual Grounding with Multi-modality LLMs
- ConcreTizer: Model Inversion Attack via Occupancy Classification and Dispersion Control for 3D Point Cloud Restoration
- MIRACLE 3D: Memory-efficient Integrated Robust Approach for Continual Learning on 3D Point Clouds via Shape Model Construction
Shape and surface completion · 10 papers
- EdgeRunner: Auto-regressive Auto-encoder for Artistic Mesh Generation (3 citations)
- Duoduo CLIP: Efficient 3D Understanding with Multi-View Images (2 citations)
- ComPC: Completing a 3D Point Cloud with 2D Diffusion Priors (1 citation)
- Implicit Neural Surface Deformation with Explicit Velocity Fields (1 citation)
- Occlusion-aware Non-Rigid Point Cloud Registration via Unsupervised Neural Deformation Correntropy (1 citation)
- Matérn Kernels for Tunable Implicit Surface Reconstruction
- Not-So-Optimal Transport Flows for 3D Point Cloud Generation
- Point-based Instance Completion with Scene Constraints
- GaussianAnything: Interactive Point Cloud Flow Matching for 3D Generation
- Shape as Line Segments: Accurate and Flexible Implicit Surface Representation