Joint Tracking and Segmentation of Multiple Targets

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

Short name:

SegTrack

Benchmark:

Description:

We propose a multi-target tracker that exploits low level image
information and associates every (super)-pixel to a specific target or
classifies it as background. As a result, we obtain a video
segmentation in addition to the classical bounding-box representation in
unconstrained, real-world videos.

Hardware:

2.7 GHz, 1 Core

Detector:

Public

Processing:

Batch

Last submitted:

April 10, 2015 (2 years ago)

Published:

April 10, 2015 at 11:33:33 CET

Submissions:

3

Open source:

Yes

Project page / code:

Reference:

A. Milan, L. Leal-Taixé, K. Schindler, I. Reid. Joint Tracking and Segmentation of Multiple Targets. In CVPR, 2015.

Benchmark performance:

MOTAMOTPFAFMTMLFPFNID Sw.FragSpecificationsDetector
22.571.71.45.8 % 63.9 % 7,89039,0206977372.7 GHz, 1 CorePublic
IDF1ID PrecisionID Recall
31.547.623.5

Detailed performance:

Sequence MOTA IDF1 MOTP FAF GT MT ML FP FN ID Sw Frag
TUD-Crossing53.949.372.80.21323.1 % 23.1 % 374561516
PETS09-S2L246.135.870.62.84226.2 % 16.7 % 1,2133,773211211
ETH-Jelmoli28.943.072.90.6458.9 % 42.2 % 2821,5091432
ETH-Linthescher11.115.173.80.21973.0 % 81.2 % 2077,7062730
ETH-Crossing23.430.774.00.1263.8 % 65.4 % 2173987
AVG-TownCentre3.38.669.30.52260.9 % 86.3 % 2356,528151108
ADL-Rundle-17.835.872.36.83225.0 % 34.4 % 3,3815,1168692
ADL-Rundle-330.834.973.91.54411.4 % 31.8 % 9146,0496869
KITTI-1640.255.273.80.7170.0 % 17.6 % 146865622
KITTI-1921.835.465.90.8623.2 % 41.9 % 8793,22279116
Venice-119.729.171.91.3170.0 % 35.3 % 5753,0573234

Raw data:


SegTrack