# Physics-Aware Diffusion for LiDAR Point Cloud Densification

arXiv: https://arxiv.org/abs/2603.26759 · HTML: https://mlsanity.com/paper/2603.26759/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 15 of its 19 references that are in the map.
- First posted on arXiv 23 Mar 2026.

## Builds on (most cited first, 15 of 15)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [nuScenes: A multimodal dataset for autonomous driving](https://mlsanity.com/paper/1903.11027/) (2019)
- [SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations](https://mlsanity.com/paper/2108.01073/) (2021)
- [Center-based 3D Object Detection and Tracking](https://mlsanity.com/paper/2006.11275/) (2020)
- [KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3D](https://mlsanity.com/paper/2109.13410/) (2021)
- [3D Shape Generation and Completion through Point-Voxel Diffusion](https://mlsanity.com/paper/2104.03670/) (2021)
- [VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Tracking](https://mlsanity.com/paper/2303.11301/) (2023)
- [In Defense of Classical Image Processing: Fast Depth Completion on the CPU](https://mlsanity.com/paper/1802.00036/) (2018)
- [Learning to Generate Realistic LiDAR Point Clouds](https://mlsanity.com/paper/2209.03954/) (2022)
- [Physics-Informed Diffusion Models](https://mlsanity.com/paper/2403.14404/) (2024)
- [Towards Realistic Scene Generation with LiDAR Diffusion Models](https://mlsanity.com/paper/2404.00815/) (2024)
- [LiDAR Data Synthesis with Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2309.09256/) (2023)
- [Scaling Diffusion Models to Real-World 3D LiDAR Scene Completion](https://mlsanity.com/paper/2403.13470/) (2024)
- [Towards foundational LiDAR world models with efficient latent flow matching](https://mlsanity.com/paper/2506.23434/) (2025)
- [FSMDet: Vision-guided feature diffusion for fully sparse 3D detector](https://mlsanity.com/paper/2409.06945/) (2024)

## Abstract (arXiv)

LiDAR perception is severely limited by the distance-dependent sparsity of distant objects. While diffusion models can recover dense geometry, they suffer from prohibitive latency and physical hallucinations manifesting as ghost points. We propose Scanline-Consistent Range-Aware Diffusion, a framework that treats densification as probabilistic refinement rather than generation. By leveraging Partial Diffusion (SDEdit) on a coarse prior, we achieve high-fidelity results in just 156ms. Our novel Ray-Consistency loss and Negative Ray Augmentation enforce sensor physics to suppress artifacts. Our method achieves state-of-the-art results on KITTI-360 and nuScenes, directly boosting off-the-shelf 3D detectors without retraining. Code will be made available.