SiameseCNN based 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:

SiameseCNN

Benchmark:

Description:

n/a

Hardware:

3 Ghz 24 cores

Detector:

Public

Processing:

Batch

Last submitted:

November 06, 2015 (1 year ago)

Published:

July 05, 2016 at 16:47:42 CET

Submissions:

1

Open source:

No

Project page / code:

n/a

Reference:

Laura Leal-Taixé, Cristian Canton-Ferrer, Konrad Schindler. Learning by Tracking: Siamese CNN for Robust Target Association. DeepVision Workshop (CVPR), Las Vegas (Nevada, USA), June 2016.

Benchmark performance:

MOTAMOTPFAFMTMLFPFNID Sw.FragSpecificationsDetector
29.071.20.98.5 % 48.4 % 5,16037,7986391,3163 Ghz 24 coresPublic
IDF1ID PrecisionID Recall
34.353.725.2

Detailed performance:

Sequence MOTA IDF1 MOTP FAF GT MT ML FP FN ID Sw Frag
TUD-Crossing73.767.473.00.41369.2 % 15.4 % 85197810
PETS09-S2L234.527.069.71.5427.1 % 19.0 % 6725,364282424
ETH-Jelmoli42.350.472.80.74524.4 % 31.1 % 3151,1193061
ETH-Linthescher16.723.274.20.11974.6 % 78.7 % 937,333927
ETH-Crossing27.532.874.10.1263.8 % 65.4 % 2070346
AVG-TownCentre19.39.169.01.62264.4 % 44.7 % 6984,927142289
ADL-Rundle-125.641.671.64.03228.1 % 28.1 % 1,9994,89633104
ADL-Rundle-339.744.072.90.34411.4 % 34.1 % 1915,9063342
KITTI-1636.937.772.60.3170.0 % 17.6 % 689812453
KITTI-1926.738.666.20.7626.5 % 29.0 % 6973,15170276
Venice-122.340.473.00.7170.0 % 41.2 % 3223,221424

Raw data:


TBX
SiameseCNN