Motion-to-Matching: A Mixed Paradigm for 3D Single Object Tracking

TL;DR

This paper introduces MTM-Tracker, a novel 3D single object tracking method that combines motion modeling and feature matching in a two-stage network to improve robustness and accuracy in LiDAR point cloud data.

Contribution

The paper proposes a mixed paradigm for 3D tracking that integrates motion priors with feature matching, addressing limitations of previous single-method approaches.

Findings

Achieves 70.9% in KITTI dataset

Achieves 51.70% in NuScenes dataset

Demonstrates competitive performance with extensive experiments

Abstract

3D single object tracking with LiDAR points is an important task in the computer vision field. Previous methods usually adopt the matching-based or motion-centric paradigms to estimate the current target status. However, the former is sensitive to the similar distractors and the sparseness of point cloud due to relying on appearance matching, while the latter usually focuses on short-term motion clues (eg. two frames) and ignores the long-term motion pattern of target. To address these issues, we propose a mixed paradigm with two stages, named MTM-Tracker, which combines motion modeling with feature matching into a single network. Specifically, in the first stage, we exploit the continuous historical boxes as motion prior and propose an encoder-decoder structure to locate target coarsely. Then, in the second stage, we introduce a feature interaction module to extract motion-aware features from consecutive point clouds and match them to refine target movement as well as regress other target states. Extensive experiments validate that our paradigm achieves competitive performance on large-scale datasets (70.9% in KITTI and 51.70% in NuScenes). The code will be open soon at https://github.com/LeoZhiheng/MTM-Tracker.git.