mfi_tst: Online multi-object tracking using multi-function integration and tracking simulation training


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:

mfi_tst

Detector:

Public

Description:

Recently, with the development of deep-learning, the performance of multi-object tracking algorithms based on the deep neural network has been greatly improved. However, most methods separate different functional modules into multiple networks and train them independently on specific tasks. When these network modules are used directly, they are not compatible with each other effectively, nor can they be better adapted to the multi-object tracking task, which leads to a poor tracking effect. Therefore, a network structure is designed to aggregate the regression of objects between frames and the extraction of appearance features into one model to improve the harmony between various functional modules of multi-object tracking. To improve the support for the multi-object tracking task, an end-to-end training method is also proposed to simulate the multi-object tracking process during the training and expand the training data by using the historical position of the target combined with the prediction of the motion model. A metric loss that can take advantage of the historical appearance features of the target is also used to train the extraction module of appearance features to improve the temporal correlation of extracted appearance features. Evaluation results on the MOTChallenge benchmark datasets show that the proposed approach achieves state-of-the-art performance.

Reference:

J. Y, H. Ge, J. Yang, Y. Tong, S. Su. Online multi-object tracking using multi-function integration and tracking simulation training. In Applied Intelligence, 2021.

Last submitted:

May 12, 2021 (3 years ago)

Published:

May 12, 2021 at 09:06:29 CET

Submissions:

1

Project page / code:

n/a

Open source:

No

Hardware:

2080TI

Runtime:

2.2 Hz

Benchmark performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT1760.158.847.2612 (26.0)699 (29.7)13,503209,47562.996.346.448.349.776.750.978.081.50.82,065 (0.0)3,829 (0.0)

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-DPM46.246.938.65101183,32248.596.441.835.944.675.737.574.579.20.32740
MOT17-01-FRCNN49.748.039.4671863,02653.194.840.738.545.874.140.772.778.40.43362
MOT17-01-SDP53.151.440.2661982,78956.894.940.340.745.473.643.272.278.20.44193
MOT17-03-DPM70.265.252.163191,58229,49771.897.948.955.851.777.858.579.782.01.1138219
MOT17-03-FRCNN71.965.852.666151,36827,87773.498.249.056.752.876.259.479.581.70.9132220
MOT17-03-SDP75.667.754.476132,77922,63078.496.749.760.053.275.863.478.381.21.9178379
MOT17-06-DPM55.557.547.057843224,81659.195.648.246.153.477.248.778.782.80.3103160
MOT17-06-FRCNN59.961.349.567564884,10965.194.049.849.653.878.653.276.882.10.4124234
MOT17-06-SDP59.861.549.978625894,02565.892.950.249.954.977.353.976.182.10.5121229
MOT17-07-DPM50.244.936.810162838,02852.596.933.840.435.876.242.277.981.40.694176
MOT17-07-FRCNN49.947.138.012144197,94453.095.536.140.438.276.942.676.781.20.893193
MOT17-07-SDP52.848.539.116145567,31356.794.536.242.838.575.745.575.880.91.1107226
MOT17-08-DPM32.435.131.1103433913,81634.695.634.827.936.779.628.879.582.90.5115155
MOT17-08-FRCNN30.833.829.593440014,11033.294.633.126.534.779.527.578.382.90.6102132
MOT17-08-SDP34.535.130.3123442813,25337.394.831.429.533.079.030.778.082.20.7152207
MOT17-12-DPM50.557.345.922301924,06353.196.050.242.253.477.644.279.883.20.23878
MOT17-12-FRCNN46.554.944.417411924,42049.095.750.339.254.476.041.080.083.50.22748
MOT17-12-SDP48.255.444.622334014,05253.292.048.241.453.573.544.376.683.00.43874
MOT17-14-DPM37.747.634.5156754610,85741.393.339.430.742.275.632.373.079.40.7104219
MOT17-14-FRCNN39.350.336.7205899910,08945.489.441.233.044.774.235.569.878.61.3139304
MOT17-14-SDP42.050.737.423521,1189,43948.989.040.435.243.773.938.169.378.41.5159381

Raw data: