Research map / Point clouds / Segmentation and detection
End-to-End Pseudo-LiDAR for Image-Based 3D Object Detection
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 |
|---|---|---|---|
| Jointly train the depth estimator and PointRCNN using detection and depth losses | Pseudo-LiDAR with PointRCNN · KITTI validation, car, moderate, IoU=0.7 | AP_3D: 50.1 -> 51.7 | Helped |
| Jointly train the depth estimator and PIXOR* using soft quantization | Pseudo-LiDAR with PIXOR* · KITTI validation, car, moderate, IoU=0.7 | AP_BEV: 61.1 -> 64.3 | Helped |
| Jointly train the depth estimator and PointRCNN using detection and depth losses | Pseudo-LiDAR with PointRCNN · KITTI test, car, moderate, IoU=0.7 | AP_3D: 42.4 -> 43.9 | Helped |
| Train Depth, RPN, and RCNN jointly rather than keeping the depth estimator fixed (ablation) | PointRCNN pseudo-LiDAR pipeline · KITTI validation, car, moderate, IoU=0.7 | AP_3D: 50.1 -> 51.7 | Helped |
| Train Depth and RCNN jointly, without the RPN loss (ablation) | PointRCNN pseudo-LiDAR pipeline · KITTI validation, car, moderate, IoU=0.7 | AP_3D: 50.1 -> 51.0 | Helped |
| Train Depth and RPN jointly, without RCNN (ablation) | PointRCNN pseudo-LiDAR pipeline · KITTI validation, car, moderate, IoU=0.7 | AP_3D: 50.1 -> 50.7 | Helped |
| Train RPN and RCNN jointly while keeping Depth fixed (ablation) | PointRCNN pseudo-LiDAR pipeline · KITTI validation, car, moderate, IoU=0.7 | AP_3D: 50.1 -> 51.2 | Helped |
| Jointly train Depth and PIXOR* Detector using soft quantization (ablation) | PIXOR* pseudo-LiDAR pipeline · KITTI validation, car, moderate, IoU=0.7 | AP_BEV: 78.3 -> 84.8 | Helped |
| Train PIXOR* Detector only, keeping Depth fixed (ablation) | PIXOR* pseudo-LiDAR pipeline · KITTI validation, car, moderate, IoU=0.7 | AP_BEV: 78.3 -> 79.2 | Helped |
| Apply soft quantization without joint end-to-end training (ablation) | PIXOR* pseudo-LiDAR pipeline · KITTI validation | AP_BEV | No clear effect |
| Train each PointRCNN subnetwork independently for more iterations rather than jointly (ablation) | PointRCNN pseudo-LiDAR pipeline · KITTI validation | AP | No clear effect |
Closest papers by meaning
- Pseudo-LiDAR From Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving (CVPR 2019 · similarity 0.94)
- Pseudo-LiDAR++: Accurate Depth for 3D Object Detection in Autonomous Driving (ICLR 2020 · similarity 0.92)
- Pseudo-Stereo for Monocular 3D Object Detection in Autonomous Driving (CVPR 2022 · similarity 0.90)
- LiDAR R-CNN: An Efficient and Universal 3D Object Detector (CVPR 2021 · similarity 0.90)
- Beyond Virtual Points: Depth-Enhanced LiDAR-only 3D Object Detection with Semi-Supervised Learning (Student Abstract) (AAAI 2025 · similarity 0.88)
- An Empirical Study of the Generalization Ability of Lidar 3D Object Detectors to Unseen Domains (CVPR 2024 · similarity 0.88)
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.