GNNMatch: Graph Convolutional Neural Networks for Multi-Object Tracking via Sinkhorn Normalization


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.

Benchmark:

Short name:

GNNMatch

Detector:

Public

Description:

Reference:

I. Papakis, A. Sarkar, A. Karpatne. GCNNMatch: Graph Convolutional Neural Networks for Multi-Object Tracking via Sinkhorn Normalization. In arXiv preprint arXiv:2010.00067, 2020.

Last submitted:

March 02, 2021 (3 years ago)

Published:

July 15, 2020 at 14:27:00 CET

Submissions:

4

Open source:

Yes

Hardware:

TITAN RTX

Runtime:

1.3 Hz

Benchmark performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT1757.356.345.4575 (24.4)787 (33.4)14,100225,04260.196.045.245.952.567.948.477.481.50.81,911 (31.8)2,837 (47.2)

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-DPM43.641.435.8512903,52945.397.037.234.544.264.635.976.980.60.21928
MOT17-01-FRCNN46.243.037.37111853,26849.394.538.136.645.460.238.774.280.20.41928
MOT17-01-SDP46.646.037.68101653,25849.595.138.337.145.962.639.074.980.10.42234
MOT17-03-DPM68.663.850.762191,90230,80370.697.548.053.853.473.056.878.481.71.3118203
MOT17-03-FRCNN69.566.652.060171,84729,97671.497.650.154.455.273.457.378.481.51.2112222
MOT17-03-SDP74.269.954.780153,11323,71977.396.351.858.057.771.761.977.080.82.1142304
MOT17-06-DPM54.233.431.861825134,77459.593.222.345.755.029.948.976.782.40.4107128
MOT17-06-FRCNN56.436.334.171608144,18364.590.324.248.254.730.552.974.181.90.7145177
MOT17-06-SDP56.234.632.274638414,16264.790.121.748.253.927.753.073.882.00.7159175
MOT17-07-DPM45.645.836.48202238,90047.397.336.836.341.861.637.877.881.50.473138
MOT17-07-FRCNN44.947.737.18173968,83547.795.338.336.441.968.138.276.381.10.883145
MOT17-07-SDP46.847.237.410154398,45450.095.137.238.042.061.740.176.281.00.995170
MOT17-08-DPM28.432.029.493719014,87629.697.035.424.540.570.225.182.484.60.35067
MOT17-08-FRCNN27.832.530.284021014,98729.196.738.124.142.474.624.782.184.70.34965
MOT17-08-SDP29.333.030.8133523214,65430.696.538.025.244.069.925.981.584.20.45785
MOT17-12-DPM46.654.044.817431084,49648.197.552.038.858.772.140.381.684.00.12030
MOT17-12-FRCNN45.051.843.715431214,63346.597.150.737.859.069.939.281.784.20.11625
MOT17-12-SDP46.354.445.117432184,42349.095.152.538.959.671.240.979.583.60.21528
MOT17-14-DPM33.638.829.6127155411,54637.592.631.428.136.959.129.572.979.70.7175222
MOT17-14-FRCNN34.139.629.9157193811,03740.388.830.729.437.355.031.569.578.91.3201269
MOT17-14-SDP36.441.431.115631,00110,52943.088.831.331.237.057.233.669.479.01.3234294

Raw data: