Method

Near-Online Multi-target Tracking [NOMT*]


Submitted on 25 Mar. 2015 00:46 by
Wongun Choi (NEC Laboratories)

Running time:0.09 s
Environment:16 cores @ 2.5 Ghz (C++)

Method Description:
* use Regionlet detetions available at:
http://www.xiaoyumu.com/s/data/kitti-tracking.zip
Parameters:
N/A
Latex Bibtex:
@article{Choi2015ICCV,
author = {Choi, Wongun},
title = {Near-Online Multi-target Tracking with
Aggregated Local Flow Descriptor
},
journal= {ICCV},
year={2015},
}

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 78.15 % 79.46 % 78.24 % 83.87 %
PEDESTRIAN 46.62 % 71.45 % 46.89 % 92.02 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 83.22 % 96.78 % 89.49 % 31854 1061 6421 9.54 % 35742 775
PEDESTRIAN 55.25 % 87.33 % 67.68 % 12874 1867 10427 16.78 % 16525 299

Benchmark MT PT ML IDS FRAG
CAR 57.23 % 29.54 % 13.23 % 31 207
PEDESTRIAN 26.12 % 39.86 % 34.02 % 63 666

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