Method

Discrete-Continuous Optimization [DCO]
http://research.milanton.net/dctracking/

Submitted on 21 May. 2014 10:21 by
Anton Milan (University of Adelaide)

Running time:0.03 s
Environment:1 core @ >3.5 Ghz (Matlab + C/C++)

Method Description:
A global multi-target tracking approach
that jointly addresses data association and
trajectory estimation by minimizing a consistent
discrete-continuous energy.
Parameters:
Default parameters with provided L-SVM detections
(threshold 0.0)
Latex Bibtex:
@inproceedings{Andriyenko2012CVPR,
Author = {Anton Andriyenko and Konrad
Schindler and Stefan Roth},
Booktitle = {CVPR},
Title = {Discrete-Continuous Optimization
for Multi-Target Tracking},
Year = {2012}
}

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 37.28 % 74.36 % 37.92 % 81.40 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 52.25 % 80.57 % 63.39 % 18486 4458 16891 40.08 % 25860 824

Benchmark MT PT ML IDS FRAG
CAR 15.54 % 53.54 % 30.92 % 220 612

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