Tracktor++v2: Tracktor++ PyTorch 1.3

MOT17-03-DPM


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.

Benchmark:

Short name:

Tracktor++v2

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 prediction 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:

November 27, 2019 (5 years ago)

Published:

March 17, 2020 at 17:34:00 CET

Submissions:

1

Open source:

Yes

Hardware:

Titan X

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
MOT1756.355.144.8498 (21.1)831 (35.3)8,866235,44958.397.445.144.948.378.447.078.681.80.51,987 (34.1)3,763 (64.6)

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-DPM41.936.232.9511233,70142.699.232.933.137.068.734.079.181.20.12535
MOT17-01-FRCNN43.839.234.36101083,49145.996.533.934.938.667.036.376.480.70.22748
MOT17-01-SDP43.838.234.16101123,47946.196.433.435.038.167.036.576.380.60.23153
MOT17-03-DPM67.961.549.860181,56831,84869.697.946.953.249.879.556.078.881.81.0169251
MOT17-03-FRCNN68.861.349.557171,63530,90870.597.845.953.848.978.456.678.681.61.1155251
MOT17-03-SDP73.365.553.070162,53425,25175.996.949.557.153.377.760.777.581.01.7199442
MOT17-06-DPM54.356.044.548851425,16356.297.944.544.649.876.246.480.883.00.182155
MOT17-06-FRCNN58.257.945.955612354,58661.196.844.347.749.874.550.179.482.50.2101222
MOT17-06-SDP58.458.146.661642864,51461.796.245.448.050.974.650.678.982.50.2101223
MOT17-07-DPM42.344.535.5521989,56043.498.737.334.138.980.735.179.882.10.292208
MOT17-07-FRCNN42.244.135.46232149,44644.197.236.934.338.480.935.678.581.90.499224
MOT17-07-SDP44.545.936.78182329,03346.597.137.736.139.779.437.578.381.70.5109279
MOT17-08-DPM26.729.927.88397615,32527.598.734.022.935.783.623.383.985.20.178106
MOT17-08-FRCNN26.330.528.484010515,40127.198.235.822.637.783.323.083.485.10.26797
MOT17-08-SDP27.531.028.5103814815,08628.697.634.523.737.181.324.282.684.60.285127
MOT17-12-DPM46.355.145.11743304,59447.099.353.238.457.279.539.583.584.70.02952
MOT17-12-FRCNN45.054.744.71545324,71345.699.253.937.258.278.938.383.384.60.02035
MOT17-12-SDP45.955.345.217441254,53347.797.153.438.457.878.040.081.484.30.12751
MOT17-14-DPM32.540.029.8118119112,14334.397.134.126.235.876.627.076.580.40.3144235
MOT17-14-FRCNN34.041.730.9137547311,55337.593.634.428.136.574.329.473.579.60.6167309
MOT17-14-SDP36.242.732.0127249911,12139.893.734.929.637.274.931.173.179.50.7180360

Raw data:


Tracktor++v2