Research map / Point clouds / Segmentation and detection
Multi-Echo LiDAR for 3D Object Detection
Yunze Man, Xinshuo Weng, Prasanna Kumar Sivakumar, Matthew O'Toole, Kris M. Kitani
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 multi-echo aggregation (MEA) over merged multi-echo points (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 50.5 -> 53.6 | Improved |
| add ambient signal to multi-echo point cloud (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 53.6 -> 53.7 | No clear effect |
| add multi-echo reflectance signal (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 53.6 -> 55.1 | Improved |
| add both ambient and ME reflectance (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 53.6 -> 55.5 | Improved |
| add MSF module to baseline (ablation) | baseline · Real LiDAR dataset, Car | overall AP: 50.5 -> 52.1 | Improved |
| add MEA module to baseline (ablation) | baseline · Real LiDAR dataset, Car | overall AP: 50.5 -> 53.6 | Improved |
| add class feature to MSF (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 54.0 -> 55.5 | Improved |
| add pixel feature to MSF (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 55.3 -> 55.5 | Improved |
| use concatenation instead of max-pooling to aggregate point-set features (ablation) | MSLiD refinement · Real LiDAR dataset, Car | overall AP: 51.1 -> 54.5 | Improved |
| use concatenation instead of mean-pooling (ablation) | MSLiD refinement · Real LiDAR dataset, Car | overall AP: 53.9 -> 54.5 | Improved |
| add point set reassignment (penetrable/impenetrable sets) (ablation) | MSLiD · Real LiDAR dataset, Car | overall AP: 54.5 -> 55.5 | Improved |
| use all echoes merged instead of only the 1st echo | SECOND · Real LiDAR dataset, Car IoU=0.7 Easy | AP: 67.9 -> 75.0 | Improved |
Closest papers by meaning
- Multimodal Virtual Point 3D Detection (NeurIPS 2021 · similarity 0.86)
- LEROjD: Lidar Extended Radar-Only Object Detection (ECCV 2024 · similarity 0.86)
- Robust 3D Object Detection using Probabilistic Point Clouds from Single-Photon LiDARs (ICCV 2025 · similarity 0.86)
- 3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection (ECCV 2020 · similarity 0.86)
- LiDAR-in-the-Loop Hyperparameter Optimization (CVPR 2023 · similarity 0.85)
- SimDistill: Simulated Multi-Modal Distillation for BEV 3D Object Detection (AAAI 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.
