Method

Multi-stage Association based on sensor fusion for 3d multi-object tracking [la][on] [MSA-MOT]


Submitted on 14 Aug. 2022 08:12 by
zhu ziming (HANGZHOU DIANZI UNIVERSITY)

Running time:0.01 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
To achieve accurate and fast
tracking, we propose a multi-stage association
framework for 3D-MOT.
Parameters:
TBD
Latex Bibtex:
@Article{s22228650,
AUTHOR = {Zhu, Ziming and Nie, Jiahao and Wu, Han
and He, Zhiwei and Gao, Mingyu},
TITLE = {MSA-MOT: Multi-Stage Association for 3D
Multimodality Multi-Object Tracking},
JOURNAL = {Sensors},
VOLUME = {22},
YEAR = {2022},
NUMBER = {22},
ARTICLE-NUMBER = {8650},
URL = {https://www.mdpi.com/1424-8220/22/22/8650},
PubMedID = {36433246},
ISSN = {1424-8220},
DOI = {10.3390/s22228650}
}

Detailed Results

From all 29 test sequences, our benchmark computes the HOTA tracking metrics (HOTA, DetA, AssA, DetRe, DetPr, AssRe, AssPr, LocA) [1] as well as the CLEARMOT, MT/PT/ML, identity switches, and fragmentation [2,3] metrics. The tables below show all of these metrics.


Benchmark HOTA DetA AssA DetRe DetPr AssRe AssPr LocA
CAR 78.52 % 75.19 % 82.56 % 82.42 % 82.21 % 85.21 % 90.16 % 87.00 %
PEDESTRIAN 44.73 % 40.93 % 49.34 % 47.09 % 55.94 % 52.88 % 65.83 % 71.21 %

Benchmark TP FP FN
CAR 32418 1974 2060
PEDESTRIAN 15389 7761 4101

Benchmark MOTA MOTP MODA IDSW sMOTA
CAR 88.01 % 85.45 % 88.27 % 91 74.29 %
PEDESTRIAN 47.86 % 64.35 % 48.76 % 209 24.16 %

Benchmark MT rate PT rate ML rate FRAG
CAR 86.77 % 12.00 % 1.23 % 428
PEDESTRIAN 33.68 % 50.17 % 16.15 % 1490

Benchmark # Dets # Tracks
CAR 34478 1097
PEDESTRIAN 19490 1133

This table as LaTeX


This figure as: png pdf

This figure as: png pdf

[1] J. Luiten, A. Os̆ep, P. Dendorfer, P. Torr, A. Geiger, L. Leal-Taixé, B. Leibe: HOTA: A Higher Order Metric for Evaluating Multi-object Tracking. IJCV 2020.
[2] K. Bernardin, R. Stiefelhagen: Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. JIVP 2008.
[3] Y. Li, C. Huang, R. Nevatia: Learning to associate: HybridBoosted multi-target tracker for crowded scene. CVPR 2009.


eXTReMe Tracker