Affinity Learning by Exploring Temporal Reinforcement within Association Chains

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

Short name:

ALExTRAC

Benchmark:

Description:

Appearance only tracking through the use of a self-supervised affinity model.

Hardware:

2.5 GHz, 4 cores

Detector:

Public

Processing:

Batch

Last submitted:

September 12, 2015 (2 years ago)

Published:

February 16, 2016 at 04:03:23 CET

Submissions:

1

Open source:

No

Project page / code:

Reference:

A. Bewley, L. Ott, F. Ramos, B. Upcroft. ALExTRAC: Affinity Learning by Exploring Temporal Reinforcement within Association Chains. In International Conference on Robotics and Automation (ICRA), (to appear) 2016.

Benchmark performance:

MOTAMOTPFAFMTMLFPFNID Sw.FragSpecificationsDetector
17.071.21.63.9 % 52.4 % 9,23339,9331,8591,8722.5 GHz, 4 coresPublic
IDF1ID PrecisionID Recall
17.325.913.0

Detailed performance:

Sequence MOTA IDF1 MOTP FAF GT MT ML FP FN ID Sw Frag
TUD-Crossing51.241.073.00.21315.4 % 15.4 % 394594050
PETS09-S2L227.512.970.61.0420.0 % 26.2 % 4496,153385359
ETH-Jelmoli32.933.073.20.64513.3 % 28.9 % 2831,3348698
ETH-Linthescher15.417.874.10.11971.5 % 76.1 % 947,36990103
ETH-Crossing19.623.174.70.1267.7 % 65.4 % 137831010
AVG-TownCentre13.321.370.01.02261.3 % 61.1 % 4425,637118152
ADL-Rundle-15.113.571.37.03215.6 % 21.9 % 3,5035,010321306
ADL-Rundle-318.116.471.83.9446.8 % 22.7 % 2,4205,523385261
KITTI-1613.912.872.10.5170.0 % 23.5 % 1051,2798073
KITTI-1913.521.066.51.1626.5 % 30.6 % 1,1283,266227326
Venice-112.512.271.91.7170.0 % 41.2 % 7573,120117134

Raw data:


TO
ALExTRAC