Research map / Detection and segmentation / 3D detection and driving
LISO: Lidar-only Self-Supervised 3D Object Detection
Stefan Andreas Baur, Frank Moosmann, Andreas Geiger
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 self-supervised scene flow (motion cues) to clustering input for Oyster (ablation) | Oyster tracker + self-training · WOD movable BEV | AP@0.4: 0.217 -> 0.255 | Improved |
| replace Oyster tracker with flow-based tracker (ablation) | LISO · WOD movable BEV | AP@0.4: 0.255 -> 0.360 | Improved |
| add self-training without tracker (ablation) | scene-flow pseudo labels · WOD movable BEV | AP@0.4: 0.177 -> 0.201 | Improved |
| add self-training with tracker (ablation) | LISO · WOD movable BEV | AP@0.4: 0.279 -> 0.360 | Improved |
| add track optimization (jerk minimization) (ablation) | LISO · WOD movable BEV | AP@0.4: 0.360 -> 0.380 | Improved |
| use ground-truth scene flow and ego-motion instead of SLIM and KISS-ICP (ablation) | LISO · WOD movable BEV | AP@0.4: 0.380 -> 0.423 | Improved |
| regenerate pseudo labels too often (s=20k steps) | iterative self-training · WOD | AP | Hurt |
| use Transfusion-L instead of CenterPoint for self-supervised training | LISO · WOD movable BEV | AP@0.4: 0.380 -> 0.327 | Hurt |
| LISO self-training vs SeMoLi on moving objects | LISO-CP · WOD moving objects 3D | AP@0.4: 0.575 -> 0.296 | Hurt |
| LISO vs Oyster on AV2 3D IoU | CenterPoint · AV2 3D IoU | AP@0.5: 0.002 -> 0.188 | Improved |
Closest papers by meaning
- 3D Object Detection with a Self-Supervised Lidar Scene Flow Backbone (ECCV 2022 · similarity 0.91)
- Image-to-Lidar Self-Supervised Distillation for Autonomous Driving Data (CVPR 2022 · similarity 0.89)
- PSA-SSL: Pose and Size-aware Self-Supervised Learning on LiDAR Point Clouds (CVPR 2025 · similarity 0.87)
- Pre-training LiDAR-based 3D Object Detectors through Colorization (ICLR 2024 · similarity 0.87)
- SeMoLi: What Moves Together Belongs Together (CVPR 2024 · similarity 0.87)
- A-Teacher: Asymmetric Network for 3D Semi-Supervised Object Detection (CVPR 2024 · 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.
