Method

TuSimple [on] [TuSimple]


Submitted on 18 Sep. 2016 07:28 by
Jian Guo (Beihang University & TuSimple)

Running time:0.6 s
Environment:1 core @ 2.5 Ghz (Matlab + C/C++)

Method Description:
Parameters:
Latex Bibtex:
@inproceedings{choi2015near,
title={Near-online multi-target tracking with aggregated local flow
descriptor},
author={Choi, Wongun},
booktitle={Proceedings of the IEEE International Conference on
Computer Vision},
pages={3029--3037},
year={2015}
}
@inproceedings{he2016deep,
title={Deep residual learning for image recognition},
author={He, Kaiming and Zhang, Xiangyu and Ren, Shaoqing and
Sun, Jian},
booktitle={Proceedings of the IEEE conference on computer vision
and pattern recognition},
pages={770--778},
year={2016}
}

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 86.62 % 83.97 % 87.48 % 87.38 %
PEDESTRIAN 58.15 % 71.93 % 58.74 % 91.37 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 90.50 % 97.99 % 94.10 % 34322 705 3602 6.34 % 39605 926
PEDESTRIAN 64.12 % 92.61 % 75.77 % 14936 1192 8359 10.72 % 18482 373

Benchmark MT PT ML IDS FRAG
CAR 72.46 % 20.77 % 6.77 % 293 501
PEDESTRIAN 30.58 % 45.36 % 24.05 % 138 818

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