Research map / 3D vision / Gaussian splatting
SplatSSC: Decoupled Depth-Guided Gaussian Splatting for Semantic Scene Completion
Rui Qian, Haozhi Cao, Tianchen Deng, Shenghai Yuan, Lihua Xie
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 |
|---|---|---|---|
| use DGA aggregator instead of GF2.agg (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 45.13 -> 48.01 | Improved |
| use DGA aggregator instead of GF.agg (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 10.45 -> 48.01 | Improved |
| add GMF depth branch fusion instead of naive depth-aware baseline (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 36.91 -> 48.01 | Improved |
| use Probability Scale Loss instead of geometry+semantic scale losses (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 46.13 -> 48.01 | Improved |
| add explicit depth loss Ld in stage 2 (ablation) | SplatSSC · Occ-ScanNet-mini | IoU (%): 60.61 -> 60.34 | Hurt |
| add semantic scale loss Lsem_scal (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 48.01 -> 48.28 | Mixed |
| increase primitives from 1200 to 4800 (scale 0.01-0.16) (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 48.87 -> 46.74 | Hurt |
| widen scale range to [0.01, 0.32] with 1200 primitives (ablation) | SplatSSC · Occ-ScanNet-mini | mIoU (%): 48.87 -> 42.38 | Hurt |
| apply GMF to frozen DAv2 (ablation) | SplatSSC · Depth branch | delta1: 0.075 -> 0.981 | Improved |
| use depth-guided initialization with 1200 primitives vs EmbodiedOcc | SplatSSC · Occ-ScanNet | mIoU (%): 45.48 -> 51.83 | Improved |
Later papers that compared with this method
2 later accepted papers report SplatSSC as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- AdaOcc: Adaptive 3D Occupancy Prediction for Embodied Tasks (NeurIPS 2026 · better on 4 of 4 shared results)
- TGSFormer: Scalable Temporal Gaussian Splatting for Embodied Semantic Scene Completion (CVPR 2026 · better on 4 of 4 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- EmbodiedOcc++ — EmbodiedOcc: Embodied 3D Occupancy Prediction for Vision-based Online Scene Understanding (ICCV 2025) · better on 8 of 8 shared results
- GaussianFormer — GaussianFormer: Scene as Gaussians for Vision-Based 3D Semantic Occupancy Prediction (ECCV 2024) · better on 2 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
- SceneSplat++: A Large Dataset and Comprehensive Benchmark for Language Gaussian Splatting (NeurIPS 2025 · similarity 0.91)
- 3D Vision-Language Gaussian Splatting (ICLR 2025 · similarity 0.90)
- Efficient Decoupled Feature 3D Gaussian Splatting via Hierarchical Compression (CVPR 2025 · similarity 0.90)
- FHGS: Feature-Homogenized Gaussian Splatting (NeurIPS 2025 · similarity 0.90)
- MonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models (CVPR 2025 · similarity 0.90)
- Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration (CVPR 2025 · similarity 0.89)
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.
