Research map / 3D vision / 3D generation
Towards Temporal Fusion Beyond the Field of View for Camera-based Semantic Scene Completion
Jongseong Bae, Junwoo Ha, Jinnyeong Heo, Yeongin Lee, Ha Young Kim
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 temporal 3D point feature alignment instead of LSS feature fusion (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 16.58 -> 18.45 | Improved |
| add historical context blurring (HCB) (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 18.45 -> 18.88 | Improved |
| add current-centric feature densification (CCFD) (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 18.45 -> 18.86 | Improved |
| add both HCB and CCFD (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 18.45 -> 19.31 | Improved |
| densification factor x2 instead of x1 (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 18.88 -> 19.31 | Improved |
| raise densification factor to x3 or x4 beyond x2 (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 19.31 -> 19.08 | Hurt |
| extend context blurring to the current frame (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 19.31 -> 18.29 | Hurt |
| increase input frames from 1 to 4 (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 16.37 -> 19.31 | Improved |
| use 5 frames instead of 4 (ablation) | C3DFusion on CGFormer · SemanticKITTI val | mIoU: 19.31 -> 18.58 | Hurt |
| replace temporal LSS fusion with C3DFusion | VoxFormer-S · VoxFormer-S, SemanticKITTI val | mIoU: 9.56 -> 15.12 | Improved |
| replace temporal LSS fusion with C3DFusion | OccFormer · OccFormer, SemanticKITTI val | mIoU: 16.91 -> 18.29 | Improved |
| replace temporal LSS fusion with C3DFusion | ScanSSC · ScanSSC, SemanticKITTI val | mIoU: 17.67 -> 18.73 | Improved |
Closest papers by meaning
- Learning Temporal 3D Semantic Scene Completion via Optical Flow Guidance (NeurIPS 2025 · similarity 0.89)
- Bi-SSC: Geometric-Semantic Bidirectional Fusion for Camera-based 3D Semantic Scene Completion (CVPR 2024 · similarity 0.89)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.88)
- VLScene: Vision-Language Guidance Distillation for Camera-Based 3D Semantic Scene Completion (AAAI 2025 · similarity 0.87)
- Memory-Augmented Re-Completion for 3D Semantic Scene Completion (AAAI 2025 · similarity 0.87)
- IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals (NeurIPS 2025 · similarity 0.86)
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.
