Method

Multi-Object Tracking and Segmentation via Neural Message Passing [MPNTrackSeg]
https://github.com/ocetintas/MPNTrackSeg

Submitted on 30 Jun. 2022 16:15 by
Orcun Cetintas (Technical University of Munich)

Running time:0.08 s
Environment:8 cores @ 2.5 Ghz (Python)

Method Description:
This work builds upon our previous CVPR 2020
(oral) paper Learning a Neural Solver for Multiple
Object Tracking and extends it by:

- integrating an attentive module to our neural
message passing scheme to yield a unified model
for multi-object tracking and segmentation
- providing an extensive evaluation of our
tracking model over three challenging datasets,
including MOT20, KITTI and the recently proposed
Human in Events dataset.
Parameters:
All hyperparameters are specified in the code.
Latex Bibtex:
@article{MPNTrackSeg,
author = {Bras{\'o}, Guillem and Cetintas,
Orcun and Leal-Taix{\'e}, Laura},
date = {2022/09/26},
doi = {10.1007/s11263-022-01678-6},
id = {Bras{\'o}2022},
isbn = {1573-1405},
journal = {International Journal of Computer
Vision},
title = {Multi-Object Tracking and
Segmentation Via Neural Message Passing},
url = {https://doi.org/10.1007/s11263-022-
01678-6},
year = {2022}}

Detailed Results

From all 29 test sequences, our benchmark computes the commonly used tracking metrics (adapted for the segmentation case): CLEARMOT, MT/PT/ML, identity switches, and fragmentations [1,2]. The tables below show all of these metrics.


Benchmark sMOTSA MOTSA MOTSP MODSA MODSP
CAR 0.00 % 0.00 % 0.00 % 0.00 % 0.00 %
PEDESTRIAN 57.30 % 77.00 % 76.00 % 77.70 % 91.90 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 0.00 % 0.00 % 0.00 % 0 0 0 0.00 % 0 0
PEDESTRIAN 81.90 % 95.20 % 88.00 % 16950 861 3747 7.80 % 23774 391

Benchmark MT PT ML IDS FRAG
CAR 0.00 % 0.00 % 0.00 % 0 0
PEDESTRIAN 56.30 % 34.10 % 9.60 % 162 620

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