Method

3D Multi-Object Tracking in Point Clouds Based on Prediction Confidence-Guided Data Association[la] [PC3T]
https://github.com/hailanyi/3D-Multi-Object-Tracker

Submitted on 24 Dec. 2020 05:37 by
hai wu (xiamen university)

Running time:0.0045 s
Environment:1 core @ >3.5 Ghz (Python + C/C++)

Method Description:
A CA motion model-based 3D multi-object tracker.
Parameters:
TBD
Latex Bibtex:
@article{wu20213d,
title={3D Multi-Object Tracking in Point Clouds
Based on Prediction Confidence-Guided Data
Association},
author={Wu, Hai and Han, Wenkai and Wen, Chenglu
and Li, Xin and Wang, Cheng},
journal={IEEE TITS},
year={2021}
}

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 88.88 % 84.37 % 89.48 % 87.52 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 92.62 % 97.75 % 95.12 % 35244 811 2807 7.29 % 41531 1024

Benchmark MT PT ML IDS FRAG
CAR 80.00 % 11.69 % 8.31 % 208 369

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