C2E: Boosting Ego-Only 3D Object Detection via Multi-Teacher Contrastive Knowledge Distillation

2026-07-02

AI Summary

This work addresses the performance bottleneck of onboard 3D object detection in complex outdoor environments, where limited field-of-view and occlusions severely degrade perception accuracy. To overcome this challenge, the authors propose a collaborative-to-ego (C2E) paradigm that transfers knowledge from multi-agent collaborative perception to a single-agent setting through a novel multi-to-single (M2S) contrastive knowledge distillation framework. This framework integrates multi-level feature enhancement, auxiliary point cloud reconstruction, and a multi-teacher contrastive distillation mechanism to effectively align discrepancies between point cloud and feature distribution domains. It seamlessly plugs into state-of-the-art 3D detectors such as CoSDH without requiring additional communication overhead. Extensive experiments on V2XSet, V2V4Real, and DAIR-V2X benchmarks demonstrate consistent improvements, with gains in 3D mAP reaching up to 8.64%.

Abstract

LiDAR-based 3D object detection is essential for autonomous driving systems. However, traditional Ego-only Perception (Eo-Perception) suffers from limited perspective and occlusions in a complex outdoor environment, leading to performance bottlenecks. Recently, research on multi-agent Collaborative Perception (Co-Perception) has demonstrated excellent performance, but high communication costs and accumulated pose error hinder its application. To address this, we explore a novel C2E (Co-Perception to Eo-Perception) paradigm through the Multi-to-Single (M2S) agent contrastive knowledge distillation framework. Our M2S framework first designs Multi-Level Feature Enhancement module to provide more stable features, and introduces Auxiliary Point Cloud Reconstruction and Multi-Teacher Contrastive Distillation mechanisms to mitigate domain gaps in point cloud and feature distributions within the C2E paradigm. Benefiting from this, our M2S can retain the excellent performance of collaborative perception while effectively avoiding the drawbacks, such as communication delays and positioning errors. Extensive experiments on the V2XSet, V2V4Real and DAIR-V2X datasets show the effectiveness and generalizability of our M2S framework when combined with the state-of-the-art CoSDH model and other excellent 3D detectors. Our M2S framework can deliver up to a 8.64% improvement in 3D mAP performance without introducing any communication costs.

Problem

• Ego-only Perception
• 3D Object Detection
• Collaborative Perception
• Communication Cost
• Occlusion

Innovation

• Contrastive Knowledge Distillation
• Multi-Teacher Distillation
• Ego-only Perception
• 3D Object Detection
• Collaborative Perception