RAR15pub: Recurrent Autoregressive Networks for Online Multi-Object Tracking

TUD-Crossing


Benchmark:

Short name:

RAR15pub

Detector:

Public

Description:

Recurrent autoregressive networks tracking framework with public detections

Reference:

K. Fang, Y. Xiang, X. Li, S. Savarese. Recurrent Autoregressive Networks for Online Multi-Object Tracking. In The IEEE Winter Conference on Applications of Computer Vision (WACV), 2018.

Last submitted:

March 27, 2017 (3 years ago)

Published:

July 23, 2018 at 09:34:33 CET

Submissions:

2

Project page / code:

Open source:

No

Hardware:

Titan X, 1.5GHZ, 1Core

Runtime:

5.4 Hz

Benchmark performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
2D MOT 201535.145.470.994 (13.0)305 (42.3)6,77132,71746.780.91.2381 (8.1)1,523 (32.6)

Detailed performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
ADL-Rundle-126.347.471.58112,1354,69049.668.44.331168
ADL-Rundle-339.643.172.24107295,34447.486.91.264148
AVG-TownCentre30.351.369.442969174,03843.577.22.025237
ETH-Crossing37.655.174.8514761938.398.20.000
ETH-Jelmoli40.658.372.79134861,00460.475.91.11879
ETH-Linthescher26.938.273.7121321996,30729.492.90.22364
KITTI-1641.257.772.50317480952.483.70.81866
KITTI-1931.350.166.13148192,80947.475.60.841300
PETS09-S2L249.636.469.6437983,92359.387.81.8135357
TUD-Crossing70.177.872.7614627375.294.70.21129
Venice-126.039.071.8184612,90136.478.31.01575

Raw data: