Method

A Joint Multi-object Tracking Method with Center-based Feature Extraction and Occlusion Handling [CJMODT-v3]


Submitted on 11 Nov. 2022 17:53 by
Junnian Li (Beijing University of Chemical Technology)

Running time:0.05 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
This work proposes a joint MOT algorithm to handle
such occlusion. Pairs of frames in complicated
environments are taken as input. A center-based
feature extraction framework is designed for
precisely detecting objects and extracting their
feature maps. A ConvGRU module is applied to learn
permanent representations by using historical
spatio-temporal information of objects. A
Hungarian matching method is applied to match the
detected objects and predicted predictions.
Parameters:
λf = 1, λs= 0.1, and λo= 1
Latex Bibtex:

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 71.09 % 73.19 % 69.63 % 75.81 % 86.95 % 71.73 % 90.06 % 87.13 %
PEDESTRIAN 41.44 % 43.54 % 39.88 % 47.98 % 68.95 % 43.47 % 73.29 % 78.27 %

Benchmark TP FP FN
CAR 29751 4641 234
PEDESTRIAN 14272 8878 1839

Benchmark MOTA MOTP MODA IDSW sMOTA
CAR 84.75 % 85.77 % 85.83 % 369 72.44 %
PEDESTRIAN 51.43 % 74.46 % 53.71 % 527 35.68 %

Benchmark MT rate PT rate ML rate FRAG
CAR 66.61 % 27.69 % 5.69 % 326
PEDESTRIAN 32.30 % 40.89 % 26.80 % 693

Benchmark # Dets # Tracks
CAR 29985 987
PEDESTRIAN 16111 877

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