Research map / 3D vision / Gaussian splatting
Discretized Gaussian Representation for Tomographic Reconstruction
Shaokai Wu, Yuxiang Lu, Yapan Guo, Wei Ji, Suizhi Huang, Fengyu Yang, Shalayiding Sirejiding, Qichen He, Jing Tong, Yanbiao Ji, Yue Ding, Hongtao Lu
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 |
|---|---|---|---|
| replace isotropic Gaussians with anisotropic Gaussians (ablation) | DGR · FIPS synthetic 50-view | PSNR: 38.74 -> 38.71 | No clear effect |
| replace isotropic Gaussians with anisotropic Gaussians (ablation) | DGR · FIPS synthetic 50-view | Time: 7m13s -> 8m05s | Hurt |
| use Fast Volume Reconstruction instead of R2-Gaussian voxelizer (ablation) | DGR · FIPS synthetic 50-view | PSNR: 38.02 -> 38.74 | Improved |
| use Fast Volume Reconstruction instead of R2-Gaussian voxelizer (ablation) | DGR · FIPS synthetic 50-view | Time: 8m37s -> 7m13s | Improved |
| add decomposition of the Mahalanobis distance (ablation) | Fast Volume Reconstruction · Per-iteration cost | Time (s): 1.05 -> 0.09 | Improved |
| increase local box size from 13 to 17 (ablation) | DGR · AAPM-Mayo 60-view | PSNR: 35.99 -> 40.25 | Improved |
| increase local box size from 17 to 19 (ablation) | DGR · AAPM-Mayo 60-view | PSNR: 40.25 -> 40.98 PSNR, 16.58 -> 26.99 min | Mixed |
| add SSIM loss to L1 loss (ablation) | DGR · FIPS 50-view cone-beam (300 iter) | PSNR: 38.66 -> 38.94 | Improved |
| add TV loss to L1 and SSIM losses (ablation) | DGR · FIPS 50-view cone-beam (300 iter) | PSNR: 38.94 -> 39.65 | Improved |
| add SSIM loss to L1 loss (ablation) | DGR · Fan-beam 60-view (300 iter) | PSNR: 38.82 -> 38.88 | Improved |
| add TV loss to L1 and SSIM losses (ablation) | DGR · Fan-beam 60-view (300 iter) | PSNR: 38.88 -> 39.15 | Improved |
| add SSIM loss to L1 loss (ablation) | DGR · Fan-beam limited-angle 90 degrees (300 iter) | PSNR: 37.69 -> 37.81 | Improved |
Closest papers by meaning
- R$^2$-Gaussian: Rectifying Radiative Gaussian Splatting for Tomographic Reconstruction (NeurIPS 2024 · similarity 0.90)
- High-fidelity 3D Object Generation from Single Image with RGBN-Volume Gaussian Reconstruction Model (CVPR 2025 · similarity 0.86)
- Volumetrically Consistent 3D Gaussian Rasterization (CVPR 2025 · similarity 0.86)
- Large Images Are Gaussians: High-Quality Large Image Representation with Levels of 2D Gaussian Splatting (AAAI 2025 · similarity 0.85)
- Learning Efficient Fuse-and-Refine for Feed-Forward 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.85)
- GVKF: Gaussian Voxel Kernel Functions for Highly Efficient Surface Reconstruction in Open Scenes (NeurIPS 2024 · 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.
