RGCN_T: tracking by a r-gcn framework

MOT17-01-DPM


Benchmark:

MOT17 |

Short name:

RGCN_T

Detector:

Public

Description:

we present a unified graph optimization framework to solve the data as- sociation problem in multiple object tracking. The proposed framework build graphs from detection nodes and formulate the graph optimization to a learnable link prediction problem. Specifically, we first use a relational graph convolution net- work (R-GCN) to learn features of both detection nodes and their linked edges. To exploit various edge information, we then use a message passing network (MPN) to aggregate dif- ferent types of edge representations and propagate them on the graph iteratively. Finally, we build an end-to-end learn- able link prediction model by using the leaned edge features to predict if two nodes should be linked. Experimental re- sults on three benchmark datasets demonstrate the effective- ness of the proposed framework, outperforming state-of-the- art methods in multi-object tracking.

Reference:

Last submitted:

October 02, 2020 (19 days ago)

Published:

October 12, 2020 at 08:40:10 CET

Submissions:

3

Project page / code:

n/a

Open source:

No

Hardware:

3 GHZ

Runtime:

59.2 Hz

Benchmark performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
MOT1763.966.179.4795 (33.8)655 (27.8)22,565179,56868.294.51.31,774 (26.0)4,182 (61.3)

Detailed performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
MOT17-01-DPM50.363.576.9811523,15451.198.40.1213
MOT17-01-FRCNN58.159.081.51053832,28964.591.60.931108
MOT17-01-SDP49.159.076.29103242,95354.291.50.7618
MOT17-03-DPM89.382.182.512703,2217,76692.696.82.1203719
MOT17-03-FRCNN70.169.178.462171,94929,29772.097.51.396176
MOT17-03-SDP75.072.177.880153,14722,94778.196.32.1119232
MOT17-06-DPM62.965.580.776482893,93466.696.40.2151271
MOT17-06-FRCNN57.058.078.985621,4323,58169.685.11.256121
MOT17-06-SDP57.759.078.995641,4683,46770.685.01.249123
MOT17-07-DPM45.350.778.09176048,57549.293.21.262117
MOT17-07-FRCNN64.564.280.22231,0504,83271.492.02.1122426
MOT17-07-SDP47.452.077.411149797,84653.690.22.069148
MOT17-08-DPM31.941.981.4163866313,66935.391.81.14656
MOT17-08-FRCNN38.040.581.1182832612,60440.396.30.5157399
MOT17-08-SDP33.542.180.9183868713,32836.991.91.14353
MOT17-12-DPM46.556.481.620413674,25750.992.30.41428
MOT17-12-FRCNN56.768.680.127205943,13563.890.30.724178
MOT17-12-SDP47.055.981.219433724,20751.592.30.41428
MOT17-14-DPM49.562.374.743451,8037,22660.986.22.4300572
MOT17-14-FRCNN34.646.775.319711,39910,58942.784.91.997183
MOT17-14-SDP37.948.375.421651,4569,91246.485.51.9113213

Raw data: