Research map / Multi-agent systems and games / Human-AI and goal-conditioned
CoLC: Communication-Efficient Collaborative Perception with LiDAR Completion
Yushan Han, Hui Zhang, Qiming Xia, Yi Jin, Yidong Li
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 pillar-level LiDAR completion and DGDA to early fusion while transmitting the full point cloud | Early Fusion (PointPillars) · OPV2V | AP@0.7: 90.69 -> 92.93 | Improved |
| CoLC at 100% LiDAR instead of early fusion on real-world data | Early Fusion (PointPillars) · DAIR-V2X | AP@0.7: 63.83 -> 62.17 | Hurt |
| transmit only half the points (CoLC*) instead of full early fusion | CoLC · V2XSim | AP@0.7: 83.61 -> 85.28 | Improved |
| use early fusion with CoLC instead of intermediate fusion under heterogeneous detectors | CoLC · Heterogeneous scenarios (V2XSim) | AP | Improved |
| replace foreground FPS with random sampling at fixed Rfg=0.2 (ablation) | FAPS · V2XSim | AP | Hurt |
| send all foreground points instead of 20% under equal bandwidth (ablation) | FAPS · V2XSim | AP | Hurt |
| send both foreground and background points instead of foreground only, foreground plus surroundings, or background only (ablation) | FAPS · V2XSim | AP | Improved |
| add CEEF completion-enhanced fusion on top of FAPS (ablation) | CoLC · V2XSim | AP | Improved |
| use VQ-based instead of MAE-based completion (ablation) | CEEF · V2XSim (Rfg=0.2, Rbg=0.1) | AP@0.7: 77.55 -> 79.28 | Improved |
| use VQ-based instead of MAE-based completion (ablation) | CEEF · V2XSim (Rfg=0.2, Rbg=0.1) | occupancy IoU: 0.633 -> 0.626 | Hurt |
| increase encoder depth from 3 to 6 layers (ablation) | completion module · V2XSim | AP@0.7: 75.93 -> 78.26 | Improved |
| increase codebook size and embedding dimension beyond the lightweight setting (ablation) | completion module · V2XSim | AP@0.7: 78.26 -> 79.28 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- Where2comm — Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps (NeurIPS 2022) · better on 2 of 3 shared results
- CoSDH — CoSDH: Communication-Efficient Collaborative Perception via Supply-Demand Awareness and Intermediate-Late Hybridization (CVPR 2025) · better on 1 of 1 shared result
- ERMVP — ERMVP: Communication-Efficient and Collaboration-Robust Multi-Vehicle Perception in Challenging Environments (CVPR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- How2comm: Communication-Efficient and Collaboration-Pragmatic Multi-Agent Perception (NeurIPS 2023 · similarity 0.89)
- INSTINCT: Instance-Level Interaction Architecture for Query-Based Collaborative Perception (ICCV 2025 · similarity 0.89)
- IFTR: An Instance-Level Fusion Transformer for Visual Collaborative Perception (ECCV 2024 · similarity 0.89)
- Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps (NeurIPS 2022 · similarity 0.89)
- C2E: Boosting Ego-Only 3D Object Detection via Multi-Teacher Contrastive Knowledge Distillation (ECCV 2026 · similarity 0.88)
- An Extensible Framework for Open Heterogeneous Collaborative Perception (ICLR 2024 · similarity 0.86)
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.
