Research map / Point clouds / Segmentation and detection
LiDAR-Based Online 3D Video Object Detection With Graph-Based Message Passing and Spatiotemporal Transformer Attention
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 |
|---|---|---|---|
| Add the full video detector with PMPNet and AST-GRU (ablation) | PointPillars · nuScenes mini train set | mAP: 21.30 -> 29.35; +8.05 | Helped |
| Replace PFN with PMPNet (ablation) | PointPillars · nuScenes mini train set | mAP: 21.30 -> 23.35; +2.05 | Helped |
| Add AST-GRU temporal aggregation (ablation) | PointPillars · nuScenes mini train set | mAP: 21.30 -> 27.28; +5.98 | Helped |
| Add vanilla ConvGRU temporal aggregation (ablation) | PointPillars · nuScenes mini train set | mAP: 21.30 -> 23.83; +2.53 | Helped |
| Add STA-GRU without TTA (ablation) | PointPillars · nuScenes mini train set | mAP: 21.30 -> 25.23; +3.93 | Helped |
| Add TTA-GRU without STA (ablation) | PointPillars · nuScenes mini train set | mAP: 21.30 -> 25.32; +4.02 | Helped |
| Use AST-GRU rather than vanilla ConvGRU (ablation) | PointPillars · nuScenes mini train set | mAP: 23.83 -> 27.28 | Helped |
| Use AST-GRU rather than temporal 3D ConvNet (ablation) | PointPillars · nuScenes mini train set | mAP: 23.12 -> 27.28 | Helped |
| Increase input length from one to two frames (ablation) | Full Model · nuScenes | mAP: 16.84 -> 19.34; +2.50 | Helped |
| Increase input length from one to three frames (ablation) | Full Model · nuScenes | mAP: 16.84 -> 20.27; +3.43 | Helped |
| Increase input length from one to four frames (ablation) | Full Model · nuScenes | mAP: 16.84 -> 20.77; +3.93 | Helped |
| Increase input length from one to five frames (ablation) | Full Model · nuScenes | mAP: 16.84 -> 21.52; +4.68 | Helped |
Closest papers by meaning
- PTT: Point-Trajectory Transformer for Efficient Temporal 3D Object Detection (CVPR 2024 · similarity 0.89)
- MGTANet: Encoding Sequential LiDAR Points Using Long Short-Term Motion-Guided Temporal Attention for 3D Object Detection (AAAI 2023 · similarity 0.89)
- PillarNeXt: Rethinking Network Designs for 3D Object Detection in LiDAR Point Clouds (CVPR 2023 · similarity 0.89)
- Joint 3D Object Detection and Tracking Using Spatio-Temporal Representation of Camera Image and LiDAR Point Clouds (AAAI 2022 · similarity 0.88)
- Not All Points Are Equal: Learning Highly Efficient Point-Based Detectors for 3D LiDAR Point Clouds (CVPR 2022 · similarity 0.87)
- GBlobs: Explicit Local Structure via Gaussian Blobs for Improved Cross-Domain LiDAR-based 3D Object Detection (CVPR 2025 · similarity 0.87)
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.