Research map / 3D vision / 3D generation
Unleashing Semantic and Geometric Priors for 3D Scene Completion
Shiyuan Chen, Wei Sui, Bohao Zhang, Zeyd Boukhers, John See, Cong Yang
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 baseline encoder with frozen Foundation Encoder (ablation) | simplified CGFormer baseline · SemanticKITTI val | mIoU: 16.53 -> 18.59 | Improved |
| add Geometry-Aware Context Adapter (ablation) | FoundationSSC (FE only) · SemanticKITTI val | mIoU: 18.59 -> 18.95 | Improved |
| add DDVM disparity-to-depth mapping (ablation) | FoundationSSC (FE only) · SemanticKITTI val | IoU: 46.61 -> 47.93 | Improved |
| add both GCA and DDVM (ablation) | FoundationSSC (FE only) · SemanticKITTI val | mIoU: 18.59 -> 19.56 | Improved |
| add Axis-Aware Fusion (ablation) | FoundationSSC (FE+GCA+DDVM) · SemanticKITTI val | mIoU: 19.56 -> 20.36 | Improved |
| use DDVM instead of CGFormer depth refinement (ablation) | FoundationSSC · SemanticKITTI val | mIoU: 19.83 -> 20.36 | Improved |
| use analytical resampling of the cost volume instead of learned DDVM (ablation) | FoundationSSC · SemanticKITTI val | mIoU: 20.36 -> 19.59 | Hurt |
| use vanilla 3D channel attention fusion instead of AAF (ablation) | FoundationSSC · SemanticKITTI val | mIoU: 20.36 -> 20.08 | Hurt |
| 3D channel attention fusion of Flss and Fvt versus Fvt alone (ablation) | FoundationSSC · SemanticKITTI val | mIoU: 19.56 -> 20.08 | Improved |
| replace EfficientNet-B7 with DINOv2-L backbone (ablation) | FoundationSSC · SemanticKITTI val | mIoU: 15.20 -> 17.79 | Improved |
| use depth-specialised DepthAnythingV2-L instead of DINOv2-L (ablation) | FoundationSSC · SemanticKITTI val | IoU: IoU 45.88 -> 46.19, mIoU 17.79 -> 17.01 | Mixed |
| use small VFM (DINOv2-S) instead of EfficientNet-B7 (ablation) | FoundationSSC · SemanticKITTI val | mIoU: 15.20 -> 15.52 | Improved |
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.
- SOAP — SOAP: Vision-Centric 3D Semantic Scene Completion with Scene-Adaptive Decoder and Occluded Region-Aware View Projection (CVPR 2025) · better on 4 of 4 shared results
- VLScene — VLScene: Vision-Language Guidance Distillation for Camera-Based 3D Semantic Scene Completion (AAAI 2025) · better on 1 of 1 shared result
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
- VLScene: Vision-Language Guidance Distillation for Camera-Based 3D Semantic Scene Completion (AAAI 2025 · similarity 0.92)
- Learning Temporal 3D Semantic Scene Completion via Optical Flow Guidance (NeurIPS 2025 · similarity 0.91)
- Three Cars Approaching within 100m! Enhancing Distant Geometry by Tri-Axis Voxel Scanning for Camera-based Semantic Scene Completion (CVPR 2025 · similarity 0.91)
- Bi-SSC: Geometric-Semantic Bidirectional Fusion for Camera-based 3D Semantic Scene Completion (CVPR 2024 · similarity 0.90)
- Memory-Augmented Re-Completion for 3D Semantic Scene Completion (AAAI 2025 · similarity 0.90)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.90)
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.
