Method

Adaptive Information Perception Paradigm for Online Multi-Object Tracking [AIPT]


Submitted on 25 Jun. 2023 06:57 by
Yukuan Zhang (Yunnan Normal University)

Running time:0.5 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
This study proposes a multi target tracking method
based on adaptive information perception paradigm,
which can improve the performance of the tracker.
Parameters:
bluure = 0.02
svt = 5
fit = 30
Latex Bibtex:

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 85.91 % 85.42 % 86.03 % 88.28 %
PEDESTRIAN 54.91 % 75.91 % 55.11 % 93.00 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 87.89 % 99.47 % 93.32 % 33574 180 4624 1.62 % 37991 866
PEDESTRIAN 58.67 % 94.58 % 72.42 % 13642 781 9610 7.02 % 15861 300

Benchmark MT PT ML IDS FRAG
CAR 66.77 % 26.62 % 6.62 % 42 460
PEDESTRIAN 23.02 % 45.36 % 31.62 % 48 743

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