A Hybrid Data Association Framework for Robust Online Multi-Object Tracking

TUD-Crossing PETS09-S2L2 ETH-Jelmoli ETH-Linthescher ETH-Crossing AVG-TownCentre ADL-Rundle-1 ADL-Rundle-3 KITTI-16 KITTI-19

Short name:

HybridDAT

Benchmark:

Description:

Global optimization algorithms have shown impressive performance in data-association based multi-object tracking, but handling online data remains a difficult hurdle to overcome. In this paper, we present a hybrid data association framework with a min-cost multi-commodity network flow for robust online
multi-object tracking.We build local target-specific models interleaved with global optimization of the optimal data association over multiple video frames. More specifically, in the mincost multi-commodity network flow, the target-specific similarities are online learned to enforce the local consistency for reducing the complexity of the global data association. Meanwhile, the
global data association taking multiple video frames into account alleviates irrecoverable errors caused by the local data association between adjacent frames. To ensure the efficiency of online tracking, we give an efficient near-optimal solution to the proposed min-cost multi-commodity flow problem, and provide the empirical proof of its sub-optimality. The comprehensive experiments on real data demonstrate the superior tracking performance of our approach in various challenging situations.

Hardware:

3.4 GHZ, 1 Core

Detector:

Public

Processing:

Online

Last submitted:

May 30, 2016 (1 year ago)

Published:

August 24, 2017 at 14:20:15 CET

Submissions:

2

Open source:

No

Project page / code:

n/a

Reference:

M. Yang, Y. Jia. A Hybrid Data Association Framework for Robust Online Multi-Object Tracking. In IEEE Transactions on Image Processing, 2016.

Benchmark performance:

MOTAMOTPFAFMTMLFPFNID Sw.FragSpecificationsDetector
35.072.61.511.4 % 42.2 % 8,45531,1403581,2673.4 GHZ, 1 CorePublic
IDF1ID PrecisionID Recall
47.761.738.9

Detailed performance:

Sequence MOTA IDF1 MOTP FAF GT MT ML FP FN ID Sw Frag
TUD-Crossing73.370.473.60.21361.5 % 15.4 % 42244834
PETS09-S2L247.746.069.31.94211.9 % 9.5 % 8174,112110349
ETH-Jelmoli41.856.973.00.64517.8 % 33.3 % 2731,1931049
ETH-Linthescher27.137.776.70.11975.1 % 70.6 % 876,4091252
ETH-Crossing24.141.474.70.1267.7 % 61.5 % 29731113
AVG-TownCentre29.244.469.01.22269.3 % 43.4 % 5324,46561246
ADL-Rundle-130.648.872.45.73228.1 % 6.3 % 2,8273,58252162
ADL-Rundle-334.143.878.42.54413.6 % 18.2 % 1,5605,0894982
KITTI-1646.365.071.70.9175.9 % 17.6 % 1987061033
KITTI-1927.553.466.31.26211.3 % 21.0 % 1,3102,53527184
Venice-137.153.674.61.71729.4 % 23.5 % 7802,0741863

Raw data:


HybridDAT