Research map / Detection and segmentation / 3D detection and driving
INS-Conv: Incremental Sparse Convolution for Online 3D Segmentation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 INS-Conv instead of standard sparse convolution on current-frustum points (ablation) | INS-Conv m64 · ScanNetv2 | mIoU: 58.0 -> 72.2 | Helped |
| use INS-Conv instead of standard sparse convolution on current-frustum points (ablation) | INS-Conv m64 · ScanNetv2 | mAP@50: 30.5 -> 60.9 | Helped |
| add neighbor propagation (ablation) | INS-Conv · ScanNetv2 | mIoU: 72.1 -> 72.2 | Helped |
| add neighbor propagation (ablation) | INS-Conv · ScanNetv2 | mAP@50: 60.9 -> 61.2 | Helped |
| add neighbor propagation (ablation) | INS-Conv · ScanNetv2 | MSE to full propagation: 9.3 -> 5.0 | Helped |
| replace neighbor propagation with full propagation (ablation) | INS-Conv · ScanNetv2 | mIoU: 72.2 -> 72.2 | No clear effect |
| add temporal consistency constraint (ablation) | INS-Conv instance embeddings · ScanNetv2 | mAP@50: 59.6 -> 61.2 | Helped |
| add uncertainty-guided input selection (ablation) | INS-Conv m64 · ScanNetv2 val | mIoU: 72.2 -> 72.4 | Helped |
| add uncertainty-guided input selection (ablation) | INS-Conv m64 · ScanNetv2 val | mAP@50: 61.2 -> 61.4 | Helped |
| add uncertainty-guided input selection (ablation) | INS-Conv m64 · scene0645_01 | average input points per frame: 19990 -> 6820 | Helped |
Closest papers by meaning
- 3D Semantic Segmentation With Submanifold Sparse Convolutional Networks (CVPR 2018 · similarity 0.88)
- OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation (CVPR 2024 · similarity 0.88)
- IINet: Implicit Intra-inter Information Fusion for Real-Time Stereo Matching (AAAI 2024 · similarity 0.85)
- OnlineAnySeg: Online Zero-Shot 3D Segmentation by Visual Foundation Model Guided 2D Mask Merging (CVPR 2025 · similarity 0.85)
- Geometry-Guided Semantic Reconstruction for 3D Instance Segmentation (NeurIPS 2026 · similarity 0.84)
- SAI3D: Segment Any Instance in 3D Scenes (CVPR 2024 · similarity 0.84)
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.