Joint Probabilistic Data Association Revisited

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

Short name:

JPDA_m

Benchmark:

Description:

In this paper, we revisit the joint probabilistic data association (JPDA) technique and propose a novel solution based on recent developments in finding the m-best solutions to an integer linear program. The key advantage of this approach is that it makes JPDA computationally tractable in applications with high target and/or clutter density, such as spot tracking in fluorescence microscopy sequences and pedestrian tracking in surveillance footage. We also show that our JPDA algorithm embedded in a simple tracking framework is surprisingly competitive with state-of-the-art global tracking methods in these two applications, while needing considerably less processing time.

Hardware:

3 GHz, 1 Core

Detector:

Public

Processing:

Batch

Last submitted:

April 14, 2015 (2 years ago)

Published:

April 22, 2015 at 09:02:53 CET

Submissions:

2

Open source:

No

Project page / code:

Reference:

H. Rezatofighi, A. Milan, Z. Zhang, Q. Shi, A. Dick, I. Reid. Joint Probabilistic Data Association Revisited. In ICCV, 2015.

Benchmark performance:

MOTAMOTPFAFMTMLFPFNID Sw.FragSpecificationsDetector
23.868.21.15.0 % 58.1 % 6,37340,0843658693 GHz, 1 CorePublic
IDF1ID PrecisionID Recall
33.854.324.5

Detailed performance:

Sequence MOTA IDF1 MOTP FAF GT MT ML FP FN ID Sw Frag
TUD-Crossing60.970.268.40.21330.8 % 23.1 % 44385226
PETS09-S2L237.634.065.92.34211.9 % 19.0 % 1,0164,858139260
ETH-Jelmoli35.649.571.50.3456.7 % 40.0 % 1451,482831
ETH-Linthescher7.012.372.80.11970.5 % 88.3 % 1128,184840
ETH-Crossing14.923.171.10.0260.0 % 76.9 % 484914
AVG-TownCentre18.331.266.80.62264.4 % 63.3 % 2585,56123108
ADL-Rundle-111.235.769.35.53215.6 % 31.3 % 2,7645,44158128
ADL-Rundle-332.332.868.01.3446.8 % 29.5 % 8016,0206470
KITTI-1637.347.972.60.6170.0 % 23.5 % 1169351633
KITTI-1931.447.165.90.6628.1 % 27.4 % 6532,98529142
Venice-115.425.871.31.0170.0 % 52.9 % 4603,3841727

Raw data:


JPDA_m