Method

Stereo3DMOT: Stereo-based 3D Multi-Object Tracking with Re-identification [Stereo3DMOT]


Submitted on 28 Feb. 2024 14:42 by
Chen Mao (University of Chinese Academy of Sciences)

Running time:0.06 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
We propose a 3D MOT system based on a cost-
effective stereo camera pair, which includes a 3D
multimodal re-identification (ReID) model capable
of multi-task learning. The ReID model obtains the
multimodal features of objects, including RGB and
point cloud information. We design data
association and trajectory management algorithms.
The data association computes an affinity matrix
for the object feature embeddings and motion
information, while the trajectory management
controls the lifecycle of the trajectories.
Parameters:
none
Latex Bibtex:
@inproceedings{mao2023stereo3dmot,
title={Stereo3DMOT: Stereo Vision Based 3D
Multi-object Tracking with Multimodal ReID},
author={Mao, Chen and Tan, Chong and Liu, Hong
and Hu, Jingqi and Zheng, Min},
booktitle={Chinese Conference on Pattern
Recognition and Computer Vision (PRCV)},
pages={495--507},
year={2023},
organization={Springer}
}

Detailed Results

From all 29 test sequences, our benchmark computes the commonly used tracking metrics CLEARMOT, MT/PT/ML, identity switches, and fragmentations [1,2]. The tables below show all of these metrics.


Benchmark MOTA MOTP MODA MODP
CAR 87.13 % 85.17 % 87.19 % 88.04 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 90.07 % 98.34 % 94.02 % 34673 584 3823 5.25 % 39374 768

Benchmark MT PT ML IDS FRAG
CAR 75.85 % 14.77 % 9.38 % 19 533

This table as LaTeX


[1] K. Bernardin, R. Stiefelhagen: Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. JIVP 2008.
[2] Y. Li, C. Huang, R. Nevatia: Learning to associate: HybridBoosted multi-target tracker for crowded scene. CVPR 2009.


eXTReMe Tracker