Tracktor++: Tracktor++


Video not available.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Rendering of new sequences is currently deactivated due to heavy load.

Short name:

Tracktor++

Detector:

Public

Description:

The problem of tracking multiple objects in a video sequence poses several challenging tasks. For tracking-by- detection these include object re-identification, motion pre- diction and dealing with occlusions. We present a tracker that accomplishes tracking without specifically targeting any of these tasks, in particular, we perform no training or optimization on tracking data. To this end, we exploit the bounding box regression of an object detector to predict the position of an object in the next frame, thereby converting a detector into a Tracktor. We demonstrate the extensibility of our Tracktor and provide a new state-of-the-art on three multi-object tracking benchmarks by extending it with a straightforward re-identification and camera motion compensation. This benchmark submission presents the results of our extended Tracktor++ multi-object tracker.

Reference:

P. Bergmann, T. Meinhardt, L. Leal-Taixé. Tracking without bells and whistles. In ICCV, 2019.

Last submitted:

April 10, 2019 (5 years ago)

Published:

April 10, 2019 at 17:29:58 CET

Submissions:

2

Open source:

Yes

Hardware:

Titan X 12 GB

Runtime:

1.5 Hz

Benchmark performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT1753.552.342.1459 (19.5)861 (36.6)12,201248,04756.096.341.742.945.475.645.077.380.90.72,072 (37.0)4,611 (82.3)

Detailed performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT17-01-DPM35.937.132.15121313,96238.695.035.929.039.173.530.174.178.80.33973
MOT17-01-FRCNN34.934.831.45104063,75341.886.932.630.737.868.932.968.577.80.93987
MOT17-01-SDP37.536.832.96102833,70642.590.735.231.238.572.033.170.478.10.64295
MOT17-03-DPM65.257.045.852191,33834,84066.798.141.151.644.775.853.979.281.60.9222343
MOT17-03-FRCNN66.459.747.455201,01433,96167.698.643.652.047.276.054.379.281.40.7189327
MOT17-03-SDP69.660.148.759162,46929,06572.296.843.555.147.574.358.077.880.81.6248525
MOT17-06-DPM52.755.744.141891845,31054.997.244.843.650.076.845.580.583.00.280168
MOT17-06-FRCNN56.759.046.251613594,64760.695.245.647.050.676.249.778.282.10.396244
MOT17-06-SDP56.859.246.558643544,63860.695.346.347.052.074.849.878.282.20.393232
MOT17-07-DPM40.542.533.56243639,60343.295.335.532.238.373.433.774.479.40.790267
MOT17-07-FRCNN39.443.133.87245559,58843.292.936.132.238.675.734.073.079.21.193258
MOT17-07-SDP41.242.633.97205969,23145.492.834.933.637.375.935.572.679.01.2111331
MOT17-08-DPM27.030.727.773821315,13028.496.633.623.035.382.223.680.283.20.383122
MOT17-08-FRCNN27.131.728.193819715,11928.496.834.722.936.981.123.579.983.10.374108
MOT17-08-SDP28.732.128.4103625314,71530.396.233.824.136.379.224.978.882.50.4103158
MOT17-12-DPM45.655.244.21544884,59647.097.952.237.556.178.738.880.983.20.12957
MOT17-12-FRCNN43.453.943.214471854,69745.895.552.036.056.277.737.878.883.20.22541
MOT17-12-SDP45.356.944.717442124,49248.295.253.037.956.579.539.778.582.90.23460
MOT17-14-DPM26.937.127.0118759112,83430.690.533.222.235.870.723.168.575.90.892285
MOT17-14-FRCNN27.138.128.412791,20212,13934.384.133.524.636.966.626.264.175.01.6132396
MOT17-14-SDP27.638.528.412791,20812,02135.084.333.024.935.669.426.664.175.01.6158434

Raw data: