Research map / Point clouds / Segmentation and detection
SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud
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 SE-SSD instead of PV-RCNN | KITTI car detection · KITTI test | 3D AP Moderate: 81.43 -> 82.54 | Helped |
| use SE-SSD instead of SA-SSD | KITTI car detection · KITTI test | 3D AP Moderate: 79.79 -> 82.54 | Helped |
| use SE-SSD instead of SA-SSD | KITTI car detection · KITTI test | inference time (ms): 40.1 -> 30.56 | Helped |
| add consistency loss to the baseline (ablation) | SE-SSD · KITTI val, car | 3D AP Moderate: 83.22 -> 84.15 | Helped |
| add ODIoU loss instead of Smooth-L1 regression (ablation) | SE-SSD · KITTI val, car | 3D AP Moderate: 83.22 -> 83.85 | Helped |
| add shape-aware data augmentation to the baseline (ablation) | SE-SSD · KITTI val, car | 3D AP Moderate: 83.22 -> 83.70 | Helped |
| add shape-aware data augmentation to consistency loss and ODIoU (ablation) | SE-SSD · KITTI val, car | 3D AP Moderate: 85.81 -> 86.12 | Helped |
| combine consistency loss with ODIoU loss (ablation) | SE-SSD · KITTI val, car | 3D AP Moderate: 84.15 -> 85.81 | Helped |
| use both confidence and box consistency terms instead of confidence-only consistency (ablation) | SE-SSD · KITTI val, car | 3D AP Moderate: 83.83 -> 84.15 | Helped |
| replace IoU-based matching with box-center distance matching (ablation) | SE-SSD consistency loss · KITTI val, car | 3D AP Moderate: 84.15 -> 80.38 | Hurt |
| filter soft targets with NMS instead of student-prediction overlap (ablation) | SE-SSD soft-target matching · KITTI val, car | 3D AP Moderate: 84.15 -> 83.49 | Hurt |
| filter soft targets by ground-truth overlap instead of student-prediction overlap (ablation) | SE-SSD soft-target matching · KITTI val, car | 3D AP Moderate: 84.15 -> 80.73 | Hurt |
Closest papers by meaning
- SESS: Self-Ensembling Semi-Supervised 3D Object Detection (CVPR 2020 · similarity 0.89)
- Structure Aware Single-Stage 3D Object Detection From Point Cloud (CVPR 2020 · similarity 0.88)
- One for All: Multi-Domain Joint Training for Point Cloud Based 3D Object Detection (NeurIPS 2024 · similarity 0.86)
- AS-Det: Active Sampling for Adaptive 3D Object Detection in Point Clouds (AAAI 2025 · similarity 0.86)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.85)
- V-DETR: DETR with Vertex Relative Position Encoding for 3D Object Detection (ICLR 2024 · similarity 0.85)
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.