Research map / 3D vision / Gaussian splatting
LiDAR-EVS: Enhance Extrapolated View Synthesis for 3D Gaussian Splatting with Pseudo-LiDAR Supervision
Yiming Huang, Xin Kang, Sipeng Zhang, Hongliang Ren, Weihua Zhang, Junjie Lai
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 pseudo extrapolated-view LiDAR supervision (both directions) (ablation) | SplatAD · Para-Lane (right extrapolated LiDAR) | Chamfer distance (m): 0.35 -> 0.18 | Improved |
| add right-shifted pseudo LiDAR on top of left-only (ablation) | Ours · Para-Lane (right extrapolated LiDAR) | Chamfer distance (m): 0.23 -> 0.18 | Improved |
| add spatially constrained dropout (rate 0.5) (ablation) | Ours · Para-Lane (right extrapolated LiDAR) | Raydrop accuracy (%): 76.77 -> 85.48 | Improved |
| raise dropout rate beyond 0.5 to 0.8 (ablation) | Ours · Para-Lane (right extrapolated LiDAR) | Depth error (m^2): 0.0012 -> 0.0039 | Hurt |
| add intensity adjustment (ablation) | Ours · PandaSet (right extrapolated LiDAR) | Chamfer distance (m): 0.42 -> 0.29 | Improved |
| add intensity adjustment (ablation) | Ours · PandaSet (right extrapolated LiDAR) | Depth error (m^2): 0.015 -> 0.006 | Improved |
| use 10 fused frames instead of 5 (ablation) | Ours · Para-Lane (right extrapolated LiDAR) | Raydrop accuracy (%): 78.9 -> 84.3 | Improved |
| use more than 10 fused frames (30) (ablation) | Ours · Para-Lane (right extrapolated LiDAR) | Raydrop accuracy (%): 84.3 -> 85.9 | No clear effect |
| apply method to NeuRAD | NeuRAD · Para-Lane (right extrapolated LiDAR) | Chamfer distance (m): 0.48 -> 0.25 | Improved |
| apply method to LiDAR-GS | LiDAR-GS · Para-Lane (right extrapolated LiDAR) | Chamfer distance (m): 0.45 -> 0.31 | Improved |
| apply method to SplatAD on interpolated views | SplatAD · Para-Lane (interpolated LiDAR) | Chamfer distance (m): 0.19 -> 0.15 | Improved |
| apply method to LiDAR-GS on interpolated views | LiDAR-GS · Para-Lane (interpolated LiDAR) | Depth error (m^2): 0.0026 -> 0.0025 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.
- SplatAD — SplatAD: Real-Time Lidar and Camera Rendering with 3D Gaussian Splatting for Autonomous Driving (CVPR 2025) · better on 11 of 12 shared results
- NeuRAD — NeuRAD: Neural Rendering for Autonomous Driving (CVPR 2024) · better on 4 of 5 shared results
- OmniRE — OmniRe: Omni Urban Scene Reconstruction (ICLR 2025) · better on 3 of 3 shared results
- LiDAR-GS — LiDAR-GS++: Improving LiDAR Gaussian Reconstruction via Diffusion Priors (AAAI 2026) · better on 1 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
- SplatAD: Real-Time Lidar and Camera Rendering with 3D Gaussian Splatting for Autonomous Driving (CVPR 2025 · similarity 0.92)
- GS-LiDAR: Generating Realistic LiDAR Point Clouds with Panoramic Gaussian Splatting (ICLR 2025 · similarity 0.91)
- LiDAR-RT: Gaussian-based Ray Tracing for Dynamic LiDAR Re-simulation (CVPR 2025 · similarity 0.89)
- Effective Multi-sensor Conditioning for Street-view Novel-view Synthesis (NeurIPS 2026 · similarity 0.89)
- NeRF-LiDAR: Generating Realistic LiDAR Point Clouds with Neural Radiance Fields (AAAI 2024 · similarity 0.88)
- DrivingGaussian: Composite Gaussian Splatting for Surrounding Dynamic Autonomous Driving Scenes (CVPR 2024 · similarity 0.88)
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.
