DCO_X: Discrete-Continuous Energy Minimization with Exclusion


Benchmark:

Short name:

DCO_X

Detector:

Public

Description:

We formulate multi-target tracking as a discrete-continuous optimization problem that handles both data association and trajectory estimation in its respective natural domain and allows leveraging powerful methods for multi-model fitting. Data association is performed using discrete optimization with label costs, yielding near optimality. Trajectory estimation is posed as a continuous fitting problem which is used in turn to update the label costs.

Project page / code:

Reference:

A. Milan, K. Schindler, S. Roth. Multi-Target Tracking by Discrete-Continuous Energy Minimization. In IEEE PAMI, 2016.

Processing:

Batch

Last submitted:

June 02, 2015 (4 years ago)

Published:

July 28, 2015 at 17:01:02 CET

Submissions:

2

Open source:

Yes

Hardware:

2.7 GHz, 1 Core

Benchmark performance:

Sequence MOTA IDF1 MOTP MT ML FP FN ID Sw.
2D MOT 201519.631.571.437.0396.010,65238,232521

Detailed performance:

Sequence MOTA IDF1 MOTP MT ML FP FN ID Sw.
ADL-Rundle-110.034.072.05.09.03,3904,92068
ADL-Rundle-316.931.672.64.011.02,6475,73069
AVG-TownCentre8.223.269.96.0157.07635,76630
ETH-Crossing16.528.375.31.018.0477874
ETH-Jelmoli30.248.871.67.016.05131,23820
ETH-Linthescher17.023.974.98.0151.02117,17924
KITTI-1634.041.273.80.02.014595820
KITTI-1917.439.365.42.018.01,3103,05547
PETS09-S2L237.527.370.72.07.06385,200189
TUD-Crossing57.355.273.72.02.04041813
Venice-113.126.571.70.05.09482,98137

Raw data:


DCO_X