Lif_TsimInt: Lifted Disjoint Paths with Application in Multiple Object Tracking

TUD-Crossing


Benchmark:

Short name:

Lif_TsimInt

Detector:

Public

Description:

Reference:

A. Hornakova, R. Henschel, B. Rosenhahn, P. Swoboda. Lifted Disjoint Paths with Application in Multiple Object Tracking. In ICML, 2020.

Last submitted:

June 24, 2020 (3 months ago)

Published:

June 26, 2020 at 11:22:20 CET

Submissions:

1

Open source:

Yes

Hardware:

3 Ghz

Runtime:

5.8 Hz

Benchmark performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
2D MOT 201547.257.675.9195 (27.0)215 (29.8)7,63524,27760.583.01.3554 (9.2)803 (13.3)

Detailed performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
ADL-Rundle-130.258.673.91523,4712,98467.964.66.942106
ADL-Rundle-358.669.281.81786213,55765.091.41.02632
AVG-TownCentre58.366.671.878442832,61963.494.10.677184
ETH-Crossing39.054.181.25114556643.690.70.213
ETH-Jelmoli55.968.980.0181238572971.382.40.9429
ETH-Linthescher34.048.879.3231164755,40039.588.10.41968
KITTI-1633.333.068.95139963362.872.81.910262
KITTI-1939.850.371.48167942,29257.179.30.812999
PETS09-S2L256.343.875.4824143,65862.193.50.9144194
TUD-Crossing87.190.873.61103910090.996.30.2311
Venice-146.261.575.3737091,73961.979.91.6715

Raw data: