Research map / 3D vision / Gaussian splatting
Joint Semantic and Rendering Enhancements in 3D Gaussian Modeling with Anisotropic Local Encoding
Jingming He, Chongyi Li, Shiqi Wang, Sam Kwong
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 anisotropic Gaussian Chebyshev descriptor (AGCD) to local encoding (ablation) | Ours · Deep Blending | mIoU: 48.3 -> 50.6 | Improved |
| add AGCD to local encoding (ablation) | Ours · Deep Blending | PSNR: 29.63 -> 29.86 | Improved |
| use transformer-based local encoding instead of MLP (ablation) | Ours · Deep Blending | mIoU: 49.2 -> 50.6 | Improved |
| add semantic guidance to adaptive Gaussian pruning (ablation) | Ours · Deep Blending | FPS: 126 -> 164 | Improved |
| add semantic guidance to adaptive Gaussian pruning (ablation) | Ours · Deep Blending | PSNR: 29.86 -> 29.86 | No clear effect |
| add adaptive densification (ablation) | Ours · Deep Blending | PSNR: 29.50 -> 29.86 | Improved |
| add adaptive densification (ablation) | Ours · Deep Blending | FPS: 173 -> 164 | Hurt |
| add semantic guidance to adaptive SH pruning (ablation) | Ours · Deep Blending | FPS: 135 -> 164 | Improved |
| enable cross-scene knowledge transfer (CSKT) | Ours · ScanNet open-vocabulary | mIoU: 64.3 -> 65.1 | Improved |
| enable CSKT | Ours · ScanNet closed-set | mIoU: 67.4 -> 68.0 | Improved |
| 3DGS-based method vs NeRF-based Panoptic Lifting on rendering | Ours · ScanNet closed-set | PSNR: 28.50 -> 27.49 | Hurt |
| Ours vs Feature 3DGS semantic segmentation | Ours · Replica | mIoU: 78.2 -> 81.1 | Improved |
Closest papers by meaning
- Tackling View-Dependent Semantics in 3D Language Gaussian Splatting (ICML 2025 · similarity 0.88)
- FHGS: Feature-Homogenized Gaussian Splatting (NeurIPS 2025 · similarity 0.88)
- Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding (AAAI 2025 · similarity 0.88)
- Efficient Decoupled Feature 3D Gaussian Splatting via Hierarchical Compression (CVPR 2025 · similarity 0.87)
- GaussianBlock: Building Part-Aware Compositional and Editable 3D Scene by Primitives and Gaussians (ICLR 2025 · similarity 0.87)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.87)
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.
