EDA_GNN: End-to-end Data Association with Graph Neural Network for Online Multiple-Object Tracking

TUD-Crossing


Benchmark:

Short name:

EDA_GNN

Detector:

Public

Description:

n/a

Reference:

Paper ID 2713

Last submitted:

November 15, 2018 (1 year ago)

Published:

November 15, 2018 at 17:07:05 CET

Submissions:

1

Project page / code:

n/a

Open source:

No

Hardware:

NVIDIA TITAN XP

Runtime:

56.4 Hz

Benchmark performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
2D MOT 201521.827.870.565 (9.0)290 (40.2)11,97034,58743.769.22.11,488 (34.0)1,851 (42.4)

Detailed performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
ADL-Rundle-111.929.271.96112,6465,44241.559.45.3115156
ADL-Rundle-324.725.771.6392,2435,24648.468.73.6166133
AVG-TownCentre13.832.768.622741,8064,00843.963.54.0347419
ETH-Crossing20.431.773.73149469330.976.70.41115
ETH-Jelmoli35.642.872.411144231,14454.976.71.06662
ETH-Linthescher20.325.873.181373306,70025.087.10.38796
KITTI-1628.526.671.40228581951.975.61.4113117
KITTI-1911.127.466.33191,6282,94844.859.51.5175293
PETS09-S2L236.523.869.1221,2544,57752.580.22.9289406
TUD-Crossing59.236.872.45210529873.088.40.54735
Venice-113.721.870.8261,1562,71240.661.62.672119

Raw data: