Research map / Point clouds / Segmentation and detection
CIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud
Read the paperCheck 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 CIA-SSD instead of the strongest listed single-stage baseline | KITTI test car detection · KITTI test | AP3D Mod (%): 79.79 -> 80.28 | Helped |
| Add global and local point-cloud augmentation (ablation) | Baseline modules · KITTI validation car detection | AP3D Mod (%): 66.39 -> 76.33 | Helped |
| Add local augmentation after global augmentation (ablation) | Baseline modules · KITTI validation car detection | AP3D Hard (%): 74.77 -> 68.64 | Hurt |
| Add ground-truth augmentation (ablation) | Baseline with global and local augmentation · KITTI validation car detection | AP3D Mod (%): 76.33 -> 78.03 | Helped |
| Train with similar-category objects (ablation) | Baseline with augmentations · KITTI validation car detection | AP3D Mod (%): 78.03 -> 78.35 | Helped |
| Filter objects outside easy, moderate, and hard difficulty levels (ablation) | Baseline trained with similar-category objects · KITTI validation car detection | AP3D Mod (%): 78.35 -> 78.73 | Helped |
| Add SSFA (ablation) | CIA-SSD baseline modules · KITTI validation car detection | AP3D Mod (%): 78.73 -> 79.17 | Helped |
| Add the confidence function to the multi-task head (ablation) | CIA-SSD with SSFA · KITTI validation car detection | AP3D Mod (%): 79.17 -> 79.63 | Helped |
| Add DI-NMS (ablation) | CIA-SSD with SSFA and confidence function · KITTI validation car detection | AP3D Mod (%): 79.63 -> 79.81 | Helped |
| Set confidence-function beta to 4 rather than 0 (ablation) | CIA-SSD confidence function · KITTI validation car detection | AP3D Mod (%): 79.30 -> 79.63 | Helped |
Closest papers by meaning
- Structure Aware Single-Stage 3D Object Detection From Point Cloud (CVPR 2020 · similarity 0.88)
- SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud (CVPR 2021 · similarity 0.87)
- AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Clouds (AAAI 2022 · similarity 0.87)
- Towards Building Self-Aware Object Detectors via Reliable Uncertainty Quantification and Calibration (CVPR 2023 · similarity 0.86)
- Accurate Single Stage Detector Using Recurrent Rolling Convolution (CVPR 2017 · similarity 0.86)
- State Space Model Meets Transformer: A New Paradigm for 3D Object Detection (ICLR 2025 · 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.