DANetwork: Deep Association Network


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.

Benchmark:

MOT16 |

Short name:

DANetwork

Detector:

Public

Description:

Deep Association Network

Reference:

C. Ma, Y. Li, F. Yang, Z. Zhang, Y. Zhuang, H. Jia, X. Xie. Deep association: End-to-end graph-based learning for multiple object tracking with conv-graph neural network. In Proceedings of the 2019 on International Conference on Multimedia Retrieval, 2019.

Last submitted:

September 19, 2023 (5 months ago)

Published:

September 19, 2023 at 19:49:48 CET

Submissions:

1

Project page / code:

n/a

Open source:

No

Hardware:

Titian Tesla K40

Runtime:

3.0 Hz

Benchmark performance:

Sequence MOTA IDF1 HOTA MT ML FP FN Rcll Prcn AssA DetA AssRe AssPr DetRe DetPr LocA FAF ID Sw. Frag
MOT1648.649.337.7100 (13.2)330 (43.5)5,85387,26052.194.237.038.841.868.041.074.079.61.0596 (0.0)804 (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
MOT16-0146.060.043.159373,41146.798.854.734.057.076.035.074.177.50.1511
MOT16-0357.351.839.841262,61041,75460.196.035.844.740.266.247.275.480.01.7244247
MOT16-0631.838.429.9171211,2226,59042.980.229.930.551.444.633.462.574.91.060166
MOT16-0745.349.936.29156338,22249.692.836.636.239.372.838.271.477.61.375147
MOT16-0837.041.934.392422610,24038.896.637.731.440.678.632.480.683.50.47070
MOT16-1239.755.142.415435214,46246.288.052.034.854.479.437.671.680.80.62233
MOT16-1428.039.128.049260412,58131.990.733.523.536.467.424.669.877.30.8120130

Raw data:


DANetwork