Learning to Divide and Conquer Tracker

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

Short name:

LDCT

Benchmark:

Description:

Test are performed on fixed camera video only:
- TUD-Crossing
- PETS09-S2L2
- AVG-TownCentre
- ADL-Rundle-3
- KITTI-16
- Venice-1

Results on other sequences are not meaningful.

Hardware:

3 GHZ, 1 Core

Detector:

Public

Processing:

Online

Last submitted:

April 22, 2015 (2 years ago)

Published:

April 22, 2015 at 22:31:13 CET

Submissions:

1

Open source:

No

Project page / code:

Reference:

F. Solera, S. Calderara, R. Cucchiara. Learning to Divide and Conquer for Online Multi-Target Tracking. In ICCV, 2015

Benchmark performance:

MOTAMOTPFAFMTMLFPFNID Sw.FragSpecificationsDetector
4.771.72.411.4 % 32.5 % 14,06632,15612,3482,9183 GHZ, 1 CorePublic
IDF1ID PrecisionID Recall
16.820.314.3

Detailed performance:

Sequence MOTA IDF1 MOTP FAF GT MT ML FP FN ID Sw Frag
TUD-Crossing67.752.582.90.51369.2 % 7.7 % 1002055130
PETS09-S2L247.426.970.82.34214.3 % 7.1 % 9953,779297300
ETH-Jelmoli-23.71.672.81.54517.8 % 24.4 % 6411,0971,399182
ETH-Linthescher-1.72.073.50.21972.0 % 67.0 % 2686,7402,076435
ETH-Crossing-6.12.573.70.42611.5 % 38.5 % 7870628044
AVG-TownCentre31.742.072.24.222615.9 % 10.6 % 1,8782,608395509
ADL-Rundle-1-68.00.370.812.73228.1 % 12.5 % 6,3584,1415,133570
ADL-Rundle-325.220.273.40.7444.5 % 61.4 % 4537,039110101
KITTI-1653.043.479.00.41711.8 % 17.6 % 916654457
KITTI-19-48.01.266.62.5624.8 % 22.6 % 2,6262,7922,490561
Venice-133.524.768.41.3170.0 % 29.4 % 5782,38473129

Raw data:


LDCT