MVCTrack: Boosting 3D Point Cloud Tracking via Multimodal-Guided Virtual Cues

TL;DR

This paper introduces MVCTrack, a novel 3D object tracking method that uses multimodal virtual cues generated from RGB data to improve tracking accuracy in sparse point cloud scenarios, demonstrating strong results on NuScenes.

Contribution

The paper proposes MVCP, a scheme that creates dense virtual cues from RGB data to enhance LiDAR-based 3D tracking, integrating multimodal information for better performance.

Findings

Achieves competitive results on NuScenes dataset.

Effectively enriches sparse point clouds with virtual cues.

Improves tracking accuracy in challenging scenarios.

Abstract

3D single object tracking is essential in autonomous driving and robotics. Existing methods often struggle with sparse and incomplete point cloud scenarios. To address these limitations, we propose a Multimodal-guided Virtual Cues Projection (MVCP) scheme that generates virtual cues to enrich sparse point clouds. Additionally, we introduce an enhanced tracker MVCTrack based on the generated virtual cues. Specifically, the MVCP scheme seamlessly integrates RGB sensors into LiDAR-based systems, leveraging a set of 2D detections to create dense 3D virtual cues that significantly improve the sparsity of point clouds. These virtual cues can naturally integrate with existing LiDAR-based 3D trackers, yielding substantial performance gains. Extensive experiments demonstrate that our method achieves competitive performance on the NuScenes dataset.